max tokens
├── .gitignore
├── LICENSE
├── README.md
├── benchmarks
├── a10g
│ ├── debug_all_results_dm.jsonl
│ └── debug_no_lora_compute_results.jsonl
├── debug_all_results_real_dm.jsonl
├── debug_bmm_all_results_synthetic_dm.jsonl
├── debug_no_lora_compute_results.jsonl
├── debug_peft_all_results_real_dm.jsonl
├── exp_suite.py
├── launch_server.py
├── paper
│ ├── ablation_cluster
│ │ ├── Ablation_study_for_adapter_cluster_size_on_A100_(40GB).pdf
│ │ ├── ablation_cluster_cv_size_1_a100_S2.jsonl
│ │ ├── ablation_cluster_cv_size_2_a100_S2.jsonl
│ │ ├── ablation_cluster_cv_size_32_a100_S2.jsonl
│ │ ├── ablation_cluster_cv_size_4_a100_S2.jsonl
│ │ ├── ablation_cluster_cv_size_8_a100_S2.jsonl
│ │ ├── ablation_cluster_size_1_a100_S2.jsonl
│ │ ├── ablation_cluster_size_1_a100_S4.jsonl
│ │ ├── ablation_cluster_size_2_a100_S2.jsonl
│ │ ├── ablation_cluster_size_2_a100_S4.jsonl
│ │ ├── ablation_cluster_size_32_a100_S2.jsonl
│ │ ├── ablation_cluster_size_32_a100_S4.jsonl
│ │ ├── ablation_cluster_size_4_a100_S2.jsonl
│ │ ├── ablation_cluster_size_4_a100_S4.jsonl
│ │ ├── ablation_cluster_size_8_a100_S2.jsonl
│ │ ├── ablation_cluster_size_8_a100_S4.jsonl
│ │ └── exp_cmds.sh
│ ├── ablation_mem
│ │ ├── ablation_bmm_a10g_S2_no_mem.jsonl
│ │ ├── ablation_mem_a10g_S1_dm.jsonl
│ │ ├── ablation_mem_a10g_S1_no_mem.jsonl
│ │ ├── ablation_mem_a10g_S2_dm.jsonl
│ │ ├── ablation_mem_a10g_S2_no_mem.jsonl
│ │ ├── exp_cmds.sh
│ │ ├── launch_server_a10g_S1_dm.sh
│ │ ├── launch_server_a10g_S1_no_mem.sh
│ │ ├── run_a10g_S1_dm.sh
│ │ ├── run_a10g_S1_no_mem.sh
│ │ └── run_no_mem_ablation.py
│ ├── ablation_merge
│ │ ├── ablation_merge_a10g_merge.jsonl
│ │ └── ablation_merge_a10g_slora.jsonl
│ ├── peft
│ │ ├── debug_peft_num_adapter_a10g_S1_peft.jsonl
│ │ ├── debug_peft_num_adapter_a10g_S2_peft.jsonl
│ │ ├── debug_peft_num_alpha_a10g_S1_peft.jsonl
│ │ ├── debug_peft_num_alpha_a10g_S2_peft.jsonl
│ │ ├── debug_peft_num_cv_a10g_S1_peft.jsonl
│ │ ├── debug_peft_num_cv_a10g_S2_peft.jsonl
│ │ ├── debug_peft_num_req_a10g_S1_peft.jsonl
│ │ ├── debug_peft_num_req_a10g_S2_peft.jsonl
│ │ ├── exp_cmds.sh
│ │ ├── exp_cmds_tab3.sh
│ │ ├── num_adapter_a100-80_S1_peft_fig3.jsonl
│ │ ├── num_adapter_a100-80_S2_peft_fig3.jsonl
│ │ └── num_adapter_a100-80_S4_peft_fig3.jsonl
│ ├── real
│ │ ├── bmm
│ │ │ ├── exp_cmds_a10g_s2.sh
│ │ │ └── real_req_rate_a10g_S2_bmm.jsonl
│ │ ├── no_mem
│ │ │ ├── exp_cmds_a10g_s2.sh
│ │ │ └── real_req_rate_a10g_S2_no_mem.jsonl
│ │ └── slora
│ │ │ ├── exp_cmds_a10g_s2.sh
│ │ │ └── real_req_rate_a10g_S2_slora.jsonl
│ └── synthetic
│ │ ├── abort
│ │ ├── backup
│ │ │ ├── synthetic_alpha_a10g_S1_abort.jsonl
│ │ │ ├── synthetic_alpha_a10g_S2_abort.jsonl
│ │ │ ├── synthetic_num_adapters_a10g_S1_abort.jsonl
│ │ │ ├── synthetic_num_adapters_a10g_S2_abort.jsonl
│ │ │ ├── synthetic_req_rate_a10g_S1_abort.jsonl
│ │ │ └── synthetic_req_rate_a10g_S2_abort.jsonl
│ │ ├── exp_cmds.sh
│ │ ├── synthetic_cv_a100-80_S4_abort.jsonl
│ │ ├── synthetic_cv_a10g_S1_abort.jsonl
│ │ └── synthetic_cv_a10g_S2_abort.jsonl
│ │ ├── bmm
│ │ ├── exp_cmds_a100-40_s4.sh
│ │ ├── exp_cmds_a100-80_s2.sh
│ │ ├── exp_cmds_a100-80_s4.sh
│ │ ├── exp_cmds_a10_s2.sh
│ │ ├── exp_cmds_a10g_s1.sh
│ │ ├── synthetic_num_adapters_a100-40_S4_bmm.jsonl
│ │ ├── synthetic_num_adapters_a100-40_S4_bmm_unused.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S2_bmm.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S4_bmm.jsonl
│ │ ├── synthetic_num_adapters_a10g_S1_bmm.jsonl
│ │ ├── synthetic_num_adapters_a10g_S2_bmm.jsonl
│ │ ├── synthetic_req_rate_a100-80_S4_bmm.jsonl
│ │ ├── synthetic_req_rate_a100-80_S4_bmm_30.jsonl
│ │ └── synthetic_req_rate_a10g_S2_bmm.jsonl
│ │ ├── lifo
│ │ ├── exp_cmds.sh
│ │ ├── synthetic_cv_a100-80_S4_lifo.jsonl
│ │ └── synthetic_cv_a10g_S2_lifo.jsonl
│ │ ├── no_mem
│ │ ├── exp_cmds_a100-40_s4.sh
│ │ ├── exp_cmds_a100-80_s2.sh
│ │ ├── exp_cmds_a100-80_s4.sh
│ │ ├── exp_cmds_a10g_S1.sh
│ │ ├── exp_cmds_a10g_S2.sh
│ │ ├── synthetic_num_adapters_a100-40_S4_no_mem.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S2_no_mem.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S4_no_mem.jsonl
│ │ ├── synthetic_num_adapters_a10g_S1_no_mem.jsonl
│ │ ├── synthetic_num_adapters_a10g_S2_no_mem.jsonl
│ │ ├── synthetic_req_rate_a100-80_S4_no_mem.jsonl
│ │ └── synthetic_req_rate_a10g_S2_no_mem.jsonl
│ │ ├── pets
│ │ ├── exp_cmds.sh
│ │ ├── synthetic_a10g_S2_pets.jsonl
│ │ ├── synthetic_alpha_a10g_S1_pets.jsonl
│ │ ├── synthetic_alpha_a10g_S2_pets.jsonl
│ │ ├── synthetic_cv_a10g_S1_pets.jsonl
│ │ ├── synthetic_cv_a10g_S2_pets.jsonl
│ │ ├── synthetic_num_adapters_a10g_S1_pets.jsonl
│ │ ├── synthetic_num_adapters_a10g_S2_pets.jsonl
│ │ ├── synthetic_req_rate_a10g_S1_pets.jsonl
│ │ └── synthetic_req_rate_a10g_S2_pets.jsonl
│ │ ├── slora
│ │ ├── backup
│ │ │ ├── synthetic_alpha_a10g_S1_slora.jsonl
│ │ │ ├── synthetic_alpha_a10g_S2_slora.jsonl
│ │ │ ├── synthetic_num_adapters_a10g_S1_slora.jsonl
│ │ │ ├── synthetic_num_adapters_a10g_S2_slora.jsonl
│ │ │ ├── synthetic_req_rate_a10g_S1_slora.jsonl
│ │ │ └── synthetic_req_rate_a10g_S2_slora.jsonl
│ │ ├── exp_cmds.sh
│ │ ├── exp_cmds_a100-40.sh
│ │ ├── exp_cmds_a100-80_S1.sh
│ │ ├── exp_cmds_a100-80_S2.sh
│ │ ├── exp_cmds_a100-80_S4.sh
│ │ ├── synthetic_cv_a100-80_S4_slora.jsonl
│ │ ├── synthetic_cv_a10g_S1_slora.jsonl
│ │ ├── synthetic_cv_a10g_S2_slora.jsonl
│ │ ├── synthetic_num_adapters_a100-40_S4_slora.jsonl
│ │ ├── synthetic_num_adapters_a100-40_S4_slora_unused.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S1_slora_table.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S2_slora.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S2_slora_table.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S4_slora.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S4_slora_table.jsonl
│ │ ├── synthetic_num_adapters_a100_80_S4_slora_large.jsonl
│ │ ├── synthetic_num_adapters_a10g_S1_slora.jsonl
│ │ ├── synthetic_num_adapters_a10g_S2_slora.jsonl
│ │ ├── synthetic_req_rate_a100-80_S4_slora.jsonl
│ │ ├── synthetic_req_rate_a10g_S1_slora.jsonl
│ │ └── synthetic_req_rate_a10g_S2_slora.jsonl
│ │ └── vllm
│ │ ├── exp_cmds_a100-80_S2.sh
│ │ ├── exp_cmds_a100-80_S3.sh
│ │ ├── exp_cmds_a10g_s2.sh
│ │ ├── synthetic_num_adapters_a100-80_S2_vllm.jsonl
│ │ ├── synthetic_num_adapters_a100-80_S3_vllm.jsonl
│ │ └── synthetic_num_adapters_a10g_S2_vllm.jsonl
├── real_trace
│ ├── clean_chat_data.py
│ └── parse_into_trace.py
├── run_exp.py
├── run_exp_peft.py
├── time_stats.py
└── trace.py
├── figures
├── memory_allocation_overview.pdf
├── overview.png
├── serving_perf.png
├── slora_tp.png
├── synthetic.png
├── tp.png
├── unifiedpaging.png
└── vllm_and_peft.png
├── setup.py
├── slora
├── common
│ ├── __init__.py
│ ├── basemodel
│ │ ├── __init__.py
│ │ ├── basemodel.py
│ │ ├── infer_struct.py
│ │ ├── layer_infer
│ │ │ ├── __init__.py
│ │ │ ├── base_layer_infer.py
│ │ │ ├── post_layer_infer.py
│ │ │ ├── pre_layer_infer.py
│ │ │ ├── template
│ │ │ │ ├── __init__.py
│ │ │ │ ├── post_layer_infer_template.py
│ │ │ │ ├── pre_layer_infer_template.py
│ │ │ │ └── transformer_layer_infer_template.py
│ │ │ └── transformer_layer_infer.py
│ │ ├── layer_weights
│ │ │ ├── __init__.py
│ │ │ ├── base_layer_weight.py
│ │ │ ├── hf_load_utils.py
│ │ │ ├── pre_and_post_layer_weight.py
│ │ │ └── transformer_layer_weight.py
│ │ └── triton_kernel
│ │ │ ├── __init__.py
│ │ │ ├── apply_penalty.py
│ │ │ ├── dequantize_gemm_int4.py
│ │ │ ├── dequantize_gemm_int8.py
│ │ │ ├── destindex_copy_kv.py
│ │ │ └── quantize_gemm_int8.py
│ ├── build_utils.py
│ ├── configs
│ │ ├── __init__.py
│ │ └── config.py
│ ├── gqa_mem_manager.py
│ ├── infer_utils.py
│ ├── int8kv_mem_manager.py
│ ├── mem_allocator.py
│ ├── mem_manager.py
│ └── ppl_int8kv_mem_manager.py
├── csrc
│ ├── bgmv
│ │ ├── bgmv_all.cu
│ │ ├── bgmv_config.h
│ │ ├── bgmv_impl.cuh
│ │ └── vec_dtypes.cuh
│ └── lora_ops.cc
├── models
│ ├── __init__.py
│ ├── bmm
│ │ └── lora_bmm_infer.py
│ ├── llama
│ │ ├── __init__.py
│ │ ├── infer_struct.py
│ │ ├── layer_infer
│ │ │ ├── __init__.py
│ │ │ ├── post_layer_infer.py
│ │ │ ├── pre_layer_infer.py
│ │ │ └── transformer_layer_infer.py
│ │ ├── layer_weights
│ │ │ ├── __init__.py
│ │ │ ├── pre_and_post_layer_weight.py
│ │ │ └── transformer_layer_weight.py
│ │ ├── model.py
│ │ └── triton_kernel
│ │ │ ├── __init__.py
│ │ │ ├── context_flashattention_nopad.py
│ │ │ ├── rmsnorm.py
│ │ │ ├── rotary_emb.py
│ │ │ ├── token_attention_nopad_att1.py
│ │ │ ├── token_attention_nopad_reduceV.py
│ │ │ ├── token_attention_nopad_softmax.py
│ │ │ └── token_attention_softmax_and_reducev.py
│ ├── llama2
│ │ ├── __init__.py
│ │ ├── layer_infer
│ │ │ ├── __init__.py
│ │ │ └── transformer_layer_infer.py
│ │ ├── layer_weights
│ │ │ ├── __init__.py
│ │ │ └── transformer_layer_weight.py
│ │ ├── model.py
│ │ └── triton_kernel
│ │ │ ├── __init__.py
│ │ │ ├── context_flashattention_nopad.py
│ │ │ ├── token_attention_nopad_att1.py
│ │ │ ├── token_attention_nopad_reduceV.py
│ │ │ ├── token_attention_nopad_softmax.py
│ │ │ └── token_attention_softmax_and_reducev.py
│ └── peft
│ │ ├── layer_weights
│ │ ├── hf_load_utils.py
│ │ └── lora_layer_weight.py
│ │ ├── lora_adapter.py
│ │ ├── lora_single_batch_infer.py
│ │ ├── lora_unordered_batch_infer.py
│ │ └── triton_kernel
│ │ ├── lora
│ │ ├── __init__.py
│ │ └── lora_prefill.py
│ │ └── tests
│ │ ├── bench_ops.py
│ │ └── benchmark_utils.py
├── mprophet
│ ├── constants.py
│ ├── hardware_parameters.py
│ ├── lora_config.py
│ ├── lora_stats.py
│ ├── measure.py
│ └── model_config.py
├── server
│ ├── __init__.py
│ ├── api_models.py
│ ├── api_server.py
│ ├── build_prompt.py
│ ├── detokenization
│ │ ├── __init__.py
│ │ ├── decode.py
│ │ └── manager.py
│ ├── httpserver
│ │ ├── __init__.py
│ │ └── manager.py
│ ├── input_params.py
│ ├── io_struct.py
│ ├── router
│ │ ├── __init__.py
│ │ ├── abort_req_queue.py
│ │ ├── cluster_req_queue.py
│ │ ├── manager.py
│ │ ├── model_infer
│ │ │ ├── __init__.py
│ │ │ ├── infer_adapter.py
│ │ │ ├── infer_batch.py
│ │ │ ├── model_rpc.py
│ │ │ ├── naive_infer_adapter.py
│ │ │ └── post_process.py
│ │ ├── peft_req_queue.py
│ │ ├── pets_req_queue.py
│ │ ├── profiler.py
│ │ ├── req_queue.py
│ │ ├── stats.py
│ │ └── vtc_req_queue.py
│ ├── sampling_params.py
│ └── tokenizer.py
└── utils
│ ├── __init__.py
│ ├── infer_utils.py
│ ├── metric.py
│ ├── model_load.py
│ ├── model_utils.py
│ └── net_utils.py
└── test
├── kernel
├── test_kernel_correctness.py
└── test_kernel_correctness_multi_rank.py
├── model
├── model_infer.py
├── model_infer_multimodal.py
├── test_llama.py
└── test_llama2.py
└── test_e2e
├── exp_suite.py
├── launch_server.py
├── run_exp.py
└── trace.py
/.gitignore:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/.gitignore
--------------------------------------------------------------------------------
/LICENSE:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/LICENSE
--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/README.md
--------------------------------------------------------------------------------
/benchmarks/a10g/debug_all_results_dm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/a10g/debug_all_results_dm.jsonl
--------------------------------------------------------------------------------
/benchmarks/a10g/debug_no_lora_compute_results.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/a10g/debug_no_lora_compute_results.jsonl
--------------------------------------------------------------------------------
/benchmarks/debug_all_results_real_dm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/debug_all_results_real_dm.jsonl
--------------------------------------------------------------------------------
/benchmarks/debug_bmm_all_results_synthetic_dm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/debug_bmm_all_results_synthetic_dm.jsonl
--------------------------------------------------------------------------------
/benchmarks/debug_no_lora_compute_results.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/debug_no_lora_compute_results.jsonl
--------------------------------------------------------------------------------
/benchmarks/debug_peft_all_results_real_dm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/debug_peft_all_results_real_dm.jsonl
--------------------------------------------------------------------------------
/benchmarks/exp_suite.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/exp_suite.py
--------------------------------------------------------------------------------
/benchmarks/launch_server.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/launch_server.py
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/Ablation_study_for_adapter_cluster_size_on_A100_(40GB).pdf:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/Ablation_study_for_adapter_cluster_size_on_A100_(40GB).pdf
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_1_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_1_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_2_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_2_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_32_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_32_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_4_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_4_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_8_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_cv_size_8_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_1_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_1_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_1_a100_S4.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_1_a100_S4.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_2_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_2_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_2_a100_S4.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_2_a100_S4.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_32_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_32_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_32_a100_S4.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_32_a100_S4.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_4_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_4_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_4_a100_S4.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_4_a100_S4.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_8_a100_S2.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_8_a100_S2.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/ablation_cluster_size_8_a100_S4.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/ablation_cluster_size_8_a100_S4.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_cluster/exp_cmds.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_cluster/exp_cmds.sh
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/ablation_bmm_a10g_S2_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/ablation_bmm_a10g_S2_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/ablation_mem_a10g_S1_dm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/ablation_mem_a10g_S1_dm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/ablation_mem_a10g_S1_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/ablation_mem_a10g_S1_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/ablation_mem_a10g_S2_dm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/ablation_mem_a10g_S2_dm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/ablation_mem_a10g_S2_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/ablation_mem_a10g_S2_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/exp_cmds.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/exp_cmds.sh
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/launch_server_a10g_S1_dm.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/launch_server_a10g_S1_dm.sh
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/launch_server_a10g_S1_no_mem.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/launch_server_a10g_S1_no_mem.sh
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/run_a10g_S1_dm.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/run_a10g_S1_dm.sh
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/run_a10g_S1_no_mem.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/run_a10g_S1_no_mem.sh
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_mem/run_no_mem_ablation.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_mem/run_no_mem_ablation.py
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_merge/ablation_merge_a10g_merge.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_merge/ablation_merge_a10g_merge.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/ablation_merge/ablation_merge_a10g_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/ablation_merge/ablation_merge_a10g_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/debug_peft_num_adapter_a10g_S1_peft.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/debug_peft_num_adapter_a10g_S1_peft.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/debug_peft_num_adapter_a10g_S2_peft.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/debug_peft_num_adapter_a10g_S2_peft.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/debug_peft_num_alpha_a10g_S1_peft.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/debug_peft_num_alpha_a10g_S1_peft.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/debug_peft_num_alpha_a10g_S2_peft.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/debug_peft_num_alpha_a10g_S2_peft.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/debug_peft_num_cv_a10g_S1_peft.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/debug_peft_num_cv_a10g_S1_peft.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/debug_peft_num_cv_a10g_S2_peft.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/debug_peft_num_cv_a10g_S2_peft.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/debug_peft_num_req_a10g_S1_peft.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/debug_peft_num_req_a10g_S1_peft.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/debug_peft_num_req_a10g_S2_peft.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/debug_peft_num_req_a10g_S2_peft.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/exp_cmds.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/exp_cmds.sh
--------------------------------------------------------------------------------
/benchmarks/paper/peft/exp_cmds_tab3.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/exp_cmds_tab3.sh
--------------------------------------------------------------------------------
/benchmarks/paper/peft/num_adapter_a100-80_S1_peft_fig3.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/num_adapter_a100-80_S1_peft_fig3.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/num_adapter_a100-80_S2_peft_fig3.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/num_adapter_a100-80_S2_peft_fig3.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/peft/num_adapter_a100-80_S4_peft_fig3.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/peft/num_adapter_a100-80_S4_peft_fig3.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/real/bmm/exp_cmds_a10g_s2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/real/bmm/exp_cmds_a10g_s2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/real/bmm/real_req_rate_a10g_S2_bmm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/real/bmm/real_req_rate_a10g_S2_bmm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/real/no_mem/exp_cmds_a10g_s2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/real/no_mem/exp_cmds_a10g_s2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/real/no_mem/real_req_rate_a10g_S2_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/real/no_mem/real_req_rate_a10g_S2_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/real/slora/exp_cmds_a10g_s2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/real/slora/exp_cmds_a10g_s2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/real/slora/real_req_rate_a10g_S2_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/real/slora/real_req_rate_a10g_S2_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/backup/synthetic_alpha_a10g_S1_abort.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/backup/synthetic_alpha_a10g_S1_abort.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/backup/synthetic_alpha_a10g_S2_abort.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/backup/synthetic_alpha_a10g_S2_abort.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/backup/synthetic_num_adapters_a10g_S1_abort.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/backup/synthetic_num_adapters_a10g_S1_abort.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/backup/synthetic_num_adapters_a10g_S2_abort.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/backup/synthetic_num_adapters_a10g_S2_abort.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/backup/synthetic_req_rate_a10g_S1_abort.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/backup/synthetic_req_rate_a10g_S1_abort.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/backup/synthetic_req_rate_a10g_S2_abort.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/backup/synthetic_req_rate_a10g_S2_abort.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/exp_cmds.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/exp_cmds.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/synthetic_cv_a100-80_S4_abort.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/synthetic_cv_a100-80_S4_abort.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/synthetic_cv_a10g_S1_abort.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/synthetic_cv_a10g_S1_abort.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/abort/synthetic_cv_a10g_S2_abort.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/abort/synthetic_cv_a10g_S2_abort.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/exp_cmds_a100-40_s4.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/exp_cmds_a100-40_s4.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/exp_cmds_a100-80_s2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/exp_cmds_a100-80_s2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/exp_cmds_a100-80_s4.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/exp_cmds_a100-80_s4.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/exp_cmds_a10_s2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/exp_cmds_a10_s2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/exp_cmds_a10g_s1.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/exp_cmds_a10g_s1.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a100-40_S4_bmm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a100-40_S4_bmm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a100-40_S4_bmm_unused.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a100-40_S4_bmm_unused.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a100-80_S2_bmm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a100-80_S2_bmm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a100-80_S4_bmm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a100-80_S4_bmm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a10g_S1_bmm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a10g_S1_bmm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a10g_S2_bmm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/synthetic_num_adapters_a10g_S2_bmm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/synthetic_req_rate_a100-80_S4_bmm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/synthetic_req_rate_a100-80_S4_bmm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/synthetic_req_rate_a100-80_S4_bmm_30.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/synthetic_req_rate_a100-80_S4_bmm_30.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/bmm/synthetic_req_rate_a10g_S2_bmm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/bmm/synthetic_req_rate_a10g_S2_bmm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/lifo/exp_cmds.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/lifo/exp_cmds.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/lifo/synthetic_cv_a100-80_S4_lifo.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/lifo/synthetic_cv_a100-80_S4_lifo.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/lifo/synthetic_cv_a10g_S2_lifo.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/lifo/synthetic_cv_a10g_S2_lifo.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/exp_cmds_a100-40_s4.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/exp_cmds_a100-40_s4.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/exp_cmds_a100-80_s2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/exp_cmds_a100-80_s2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/exp_cmds_a100-80_s4.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/exp_cmds_a100-80_s4.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/exp_cmds_a10g_S1.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/exp_cmds_a10g_S1.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/exp_cmds_a10g_S2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/exp_cmds_a10g_S2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a100-40_S4_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a100-40_S4_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a100-80_S2_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a100-80_S2_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a100-80_S4_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a100-80_S4_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a10g_S1_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a10g_S1_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a10g_S2_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/synthetic_num_adapters_a10g_S2_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/synthetic_req_rate_a100-80_S4_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/synthetic_req_rate_a100-80_S4_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/no_mem/synthetic_req_rate_a10g_S2_no_mem.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/no_mem/synthetic_req_rate_a10g_S2_no_mem.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/exp_cmds.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/exp_cmds.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/synthetic_a10g_S2_pets.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/synthetic_a10g_S2_pets.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/synthetic_alpha_a10g_S1_pets.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/synthetic_alpha_a10g_S1_pets.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/synthetic_alpha_a10g_S2_pets.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/synthetic_alpha_a10g_S2_pets.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/synthetic_cv_a10g_S1_pets.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/synthetic_cv_a10g_S1_pets.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/synthetic_cv_a10g_S2_pets.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/synthetic_cv_a10g_S2_pets.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/synthetic_num_adapters_a10g_S1_pets.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/synthetic_num_adapters_a10g_S1_pets.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/synthetic_num_adapters_a10g_S2_pets.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/synthetic_num_adapters_a10g_S2_pets.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/synthetic_req_rate_a10g_S1_pets.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/synthetic_req_rate_a10g_S1_pets.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/pets/synthetic_req_rate_a10g_S2_pets.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/pets/synthetic_req_rate_a10g_S2_pets.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/backup/synthetic_alpha_a10g_S1_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/backup/synthetic_alpha_a10g_S1_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/backup/synthetic_alpha_a10g_S2_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/backup/synthetic_alpha_a10g_S2_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/backup/synthetic_num_adapters_a10g_S1_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/backup/synthetic_num_adapters_a10g_S1_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/backup/synthetic_num_adapters_a10g_S2_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/backup/synthetic_num_adapters_a10g_S2_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/backup/synthetic_req_rate_a10g_S1_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/backup/synthetic_req_rate_a10g_S1_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/backup/synthetic_req_rate_a10g_S2_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/backup/synthetic_req_rate_a10g_S2_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/exp_cmds.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/exp_cmds.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/exp_cmds_a100-40.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/exp_cmds_a100-40.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/exp_cmds_a100-80_S1.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/exp_cmds_a100-80_S1.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/exp_cmds_a100-80_S2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/exp_cmds_a100-80_S2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/exp_cmds_a100-80_S4.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/exp_cmds_a100-80_S4.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_cv_a100-80_S4_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_cv_a100-80_S4_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_cv_a10g_S1_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_cv_a10g_S1_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_cv_a10g_S2_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_cv_a10g_S2_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-40_S4_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-40_S4_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-40_S4_slora_unused.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-40_S4_slora_unused.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S1_slora_table.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S1_slora_table.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S2_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S2_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S2_slora_table.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S2_slora_table.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S4_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S4_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S4_slora_table.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100-80_S4_slora_table.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100_80_S4_slora_large.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a100_80_S4_slora_large.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a10g_S1_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a10g_S1_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a10g_S2_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_num_adapters_a10g_S2_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_req_rate_a100-80_S4_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_req_rate_a100-80_S4_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_req_rate_a10g_S1_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_req_rate_a10g_S1_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/slora/synthetic_req_rate_a10g_S2_slora.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/slora/synthetic_req_rate_a10g_S2_slora.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/vllm/exp_cmds_a100-80_S2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/vllm/exp_cmds_a100-80_S2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/vllm/exp_cmds_a100-80_S3.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/vllm/exp_cmds_a100-80_S3.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/vllm/exp_cmds_a10g_s2.sh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/vllm/exp_cmds_a10g_s2.sh
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/vllm/synthetic_num_adapters_a100-80_S2_vllm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/vllm/synthetic_num_adapters_a100-80_S2_vllm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/vllm/synthetic_num_adapters_a100-80_S3_vllm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/vllm/synthetic_num_adapters_a100-80_S3_vllm.jsonl
--------------------------------------------------------------------------------
/benchmarks/paper/synthetic/vllm/synthetic_num_adapters_a10g_S2_vllm.jsonl:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/paper/synthetic/vllm/synthetic_num_adapters_a10g_S2_vllm.jsonl
--------------------------------------------------------------------------------
/benchmarks/real_trace/clean_chat_data.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/real_trace/clean_chat_data.py
--------------------------------------------------------------------------------
/benchmarks/real_trace/parse_into_trace.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/real_trace/parse_into_trace.py
--------------------------------------------------------------------------------
/benchmarks/run_exp.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/run_exp.py
--------------------------------------------------------------------------------
/benchmarks/run_exp_peft.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/run_exp_peft.py
--------------------------------------------------------------------------------
/benchmarks/time_stats.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/time_stats.py
--------------------------------------------------------------------------------
/benchmarks/trace.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/benchmarks/trace.py
--------------------------------------------------------------------------------
/figures/memory_allocation_overview.pdf:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/figures/memory_allocation_overview.pdf
--------------------------------------------------------------------------------
/figures/overview.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/figures/overview.png
--------------------------------------------------------------------------------
/figures/serving_perf.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/figures/serving_perf.png
--------------------------------------------------------------------------------
/figures/slora_tp.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/figures/slora_tp.png
--------------------------------------------------------------------------------
/figures/synthetic.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/figures/synthetic.png
--------------------------------------------------------------------------------
/figures/tp.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/figures/tp.png
--------------------------------------------------------------------------------
/figures/unifiedpaging.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/figures/unifiedpaging.png
--------------------------------------------------------------------------------
/figures/vllm_and_peft.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/figures/vllm_and_peft.png
--------------------------------------------------------------------------------
/setup.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/setup.py
--------------------------------------------------------------------------------
/slora/common/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/common/basemodel/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/__init__.py
--------------------------------------------------------------------------------
/slora/common/basemodel/basemodel.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/basemodel.py
--------------------------------------------------------------------------------
/slora/common/basemodel/infer_struct.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/infer_struct.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_infer/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_infer/base_layer_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_infer/base_layer_infer.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_infer/post_layer_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_infer/post_layer_infer.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_infer/pre_layer_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_infer/pre_layer_infer.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_infer/template/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_infer/template/post_layer_infer_template.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_infer/template/post_layer_infer_template.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_infer/template/pre_layer_infer_template.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_infer/template/pre_layer_infer_template.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_infer/template/transformer_layer_infer_template.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_infer/template/transformer_layer_infer_template.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_infer/transformer_layer_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_infer/transformer_layer_infer.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_weights/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_weights/base_layer_weight.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_weights/base_layer_weight.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_weights/hf_load_utils.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_weights/hf_load_utils.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_weights/pre_and_post_layer_weight.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_weights/pre_and_post_layer_weight.py
--------------------------------------------------------------------------------
/slora/common/basemodel/layer_weights/transformer_layer_weight.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/layer_weights/transformer_layer_weight.py
--------------------------------------------------------------------------------
/slora/common/basemodel/triton_kernel/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/common/basemodel/triton_kernel/apply_penalty.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/triton_kernel/apply_penalty.py
--------------------------------------------------------------------------------
/slora/common/basemodel/triton_kernel/dequantize_gemm_int4.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/triton_kernel/dequantize_gemm_int4.py
--------------------------------------------------------------------------------
/slora/common/basemodel/triton_kernel/dequantize_gemm_int8.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/triton_kernel/dequantize_gemm_int8.py
--------------------------------------------------------------------------------
/slora/common/basemodel/triton_kernel/destindex_copy_kv.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/triton_kernel/destindex_copy_kv.py
--------------------------------------------------------------------------------
/slora/common/basemodel/triton_kernel/quantize_gemm_int8.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/basemodel/triton_kernel/quantize_gemm_int8.py
--------------------------------------------------------------------------------
/slora/common/build_utils.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/build_utils.py
--------------------------------------------------------------------------------
/slora/common/configs/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/common/configs/config.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/configs/config.py
--------------------------------------------------------------------------------
/slora/common/gqa_mem_manager.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/gqa_mem_manager.py
--------------------------------------------------------------------------------
/slora/common/infer_utils.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/infer_utils.py
--------------------------------------------------------------------------------
/slora/common/int8kv_mem_manager.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/int8kv_mem_manager.py
--------------------------------------------------------------------------------
/slora/common/mem_allocator.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/mem_allocator.py
--------------------------------------------------------------------------------
/slora/common/mem_manager.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/mem_manager.py
--------------------------------------------------------------------------------
/slora/common/ppl_int8kv_mem_manager.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/common/ppl_int8kv_mem_manager.py
--------------------------------------------------------------------------------
/slora/csrc/bgmv/bgmv_all.cu:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/csrc/bgmv/bgmv_all.cu
--------------------------------------------------------------------------------
/slora/csrc/bgmv/bgmv_config.h:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/csrc/bgmv/bgmv_config.h
--------------------------------------------------------------------------------
/slora/csrc/bgmv/bgmv_impl.cuh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/csrc/bgmv/bgmv_impl.cuh
--------------------------------------------------------------------------------
/slora/csrc/bgmv/vec_dtypes.cuh:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/csrc/bgmv/vec_dtypes.cuh
--------------------------------------------------------------------------------
/slora/csrc/lora_ops.cc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/csrc/lora_ops.cc
--------------------------------------------------------------------------------
/slora/models/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/bmm/lora_bmm_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/bmm/lora_bmm_infer.py
--------------------------------------------------------------------------------
/slora/models/llama/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/llama/infer_struct.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/infer_struct.py
--------------------------------------------------------------------------------
/slora/models/llama/layer_infer/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/llama/layer_infer/post_layer_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/layer_infer/post_layer_infer.py
--------------------------------------------------------------------------------
/slora/models/llama/layer_infer/pre_layer_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/layer_infer/pre_layer_infer.py
--------------------------------------------------------------------------------
/slora/models/llama/layer_infer/transformer_layer_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/layer_infer/transformer_layer_infer.py
--------------------------------------------------------------------------------
/slora/models/llama/layer_weights/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/llama/layer_weights/pre_and_post_layer_weight.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/layer_weights/pre_and_post_layer_weight.py
--------------------------------------------------------------------------------
/slora/models/llama/layer_weights/transformer_layer_weight.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/layer_weights/transformer_layer_weight.py
--------------------------------------------------------------------------------
/slora/models/llama/model.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/model.py
--------------------------------------------------------------------------------
/slora/models/llama/triton_kernel/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/llama/triton_kernel/context_flashattention_nopad.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/triton_kernel/context_flashattention_nopad.py
--------------------------------------------------------------------------------
/slora/models/llama/triton_kernel/rmsnorm.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/triton_kernel/rmsnorm.py
--------------------------------------------------------------------------------
/slora/models/llama/triton_kernel/rotary_emb.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/triton_kernel/rotary_emb.py
--------------------------------------------------------------------------------
/slora/models/llama/triton_kernel/token_attention_nopad_att1.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/triton_kernel/token_attention_nopad_att1.py
--------------------------------------------------------------------------------
/slora/models/llama/triton_kernel/token_attention_nopad_reduceV.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/triton_kernel/token_attention_nopad_reduceV.py
--------------------------------------------------------------------------------
/slora/models/llama/triton_kernel/token_attention_nopad_softmax.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/triton_kernel/token_attention_nopad_softmax.py
--------------------------------------------------------------------------------
/slora/models/llama/triton_kernel/token_attention_softmax_and_reducev.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama/triton_kernel/token_attention_softmax_and_reducev.py
--------------------------------------------------------------------------------
/slora/models/llama2/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/llama2/layer_infer/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/llama2/layer_infer/transformer_layer_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama2/layer_infer/transformer_layer_infer.py
--------------------------------------------------------------------------------
/slora/models/llama2/layer_weights/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/llama2/layer_weights/transformer_layer_weight.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama2/layer_weights/transformer_layer_weight.py
--------------------------------------------------------------------------------
/slora/models/llama2/model.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama2/model.py
--------------------------------------------------------------------------------
/slora/models/llama2/triton_kernel/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/llama2/triton_kernel/context_flashattention_nopad.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama2/triton_kernel/context_flashattention_nopad.py
--------------------------------------------------------------------------------
/slora/models/llama2/triton_kernel/token_attention_nopad_att1.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama2/triton_kernel/token_attention_nopad_att1.py
--------------------------------------------------------------------------------
/slora/models/llama2/triton_kernel/token_attention_nopad_reduceV.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama2/triton_kernel/token_attention_nopad_reduceV.py
--------------------------------------------------------------------------------
/slora/models/llama2/triton_kernel/token_attention_nopad_softmax.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama2/triton_kernel/token_attention_nopad_softmax.py
--------------------------------------------------------------------------------
/slora/models/llama2/triton_kernel/token_attention_softmax_and_reducev.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/llama2/triton_kernel/token_attention_softmax_and_reducev.py
--------------------------------------------------------------------------------
/slora/models/peft/layer_weights/hf_load_utils.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/peft/layer_weights/hf_load_utils.py
--------------------------------------------------------------------------------
/slora/models/peft/layer_weights/lora_layer_weight.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/peft/layer_weights/lora_layer_weight.py
--------------------------------------------------------------------------------
/slora/models/peft/lora_adapter.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/peft/lora_adapter.py
--------------------------------------------------------------------------------
/slora/models/peft/lora_single_batch_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/peft/lora_single_batch_infer.py
--------------------------------------------------------------------------------
/slora/models/peft/lora_unordered_batch_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/peft/lora_unordered_batch_infer.py
--------------------------------------------------------------------------------
/slora/models/peft/triton_kernel/lora/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/models/peft/triton_kernel/lora/lora_prefill.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/peft/triton_kernel/lora/lora_prefill.py
--------------------------------------------------------------------------------
/slora/models/peft/triton_kernel/tests/bench_ops.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/peft/triton_kernel/tests/bench_ops.py
--------------------------------------------------------------------------------
/slora/models/peft/triton_kernel/tests/benchmark_utils.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/models/peft/triton_kernel/tests/benchmark_utils.py
--------------------------------------------------------------------------------
/slora/mprophet/constants.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/mprophet/constants.py
--------------------------------------------------------------------------------
/slora/mprophet/hardware_parameters.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/mprophet/hardware_parameters.py
--------------------------------------------------------------------------------
/slora/mprophet/lora_config.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/mprophet/lora_config.py
--------------------------------------------------------------------------------
/slora/mprophet/lora_stats.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/mprophet/lora_stats.py
--------------------------------------------------------------------------------
/slora/mprophet/measure.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/mprophet/measure.py
--------------------------------------------------------------------------------
/slora/mprophet/model_config.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/mprophet/model_config.py
--------------------------------------------------------------------------------
/slora/server/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/server/api_models.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/api_models.py
--------------------------------------------------------------------------------
/slora/server/api_server.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/api_server.py
--------------------------------------------------------------------------------
/slora/server/build_prompt.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/build_prompt.py
--------------------------------------------------------------------------------
/slora/server/detokenization/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/server/detokenization/decode.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/detokenization/decode.py
--------------------------------------------------------------------------------
/slora/server/detokenization/manager.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/detokenization/manager.py
--------------------------------------------------------------------------------
/slora/server/httpserver/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/server/httpserver/manager.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/httpserver/manager.py
--------------------------------------------------------------------------------
/slora/server/input_params.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/input_params.py
--------------------------------------------------------------------------------
/slora/server/io_struct.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/io_struct.py
--------------------------------------------------------------------------------
/slora/server/router/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/server/router/abort_req_queue.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/abort_req_queue.py
--------------------------------------------------------------------------------
/slora/server/router/cluster_req_queue.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/cluster_req_queue.py
--------------------------------------------------------------------------------
/slora/server/router/manager.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/manager.py
--------------------------------------------------------------------------------
/slora/server/router/model_infer/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/server/router/model_infer/infer_adapter.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/model_infer/infer_adapter.py
--------------------------------------------------------------------------------
/slora/server/router/model_infer/infer_batch.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/model_infer/infer_batch.py
--------------------------------------------------------------------------------
/slora/server/router/model_infer/model_rpc.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/model_infer/model_rpc.py
--------------------------------------------------------------------------------
/slora/server/router/model_infer/naive_infer_adapter.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/model_infer/naive_infer_adapter.py
--------------------------------------------------------------------------------
/slora/server/router/model_infer/post_process.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/model_infer/post_process.py
--------------------------------------------------------------------------------
/slora/server/router/peft_req_queue.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/peft_req_queue.py
--------------------------------------------------------------------------------
/slora/server/router/pets_req_queue.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/pets_req_queue.py
--------------------------------------------------------------------------------
/slora/server/router/profiler.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/profiler.py
--------------------------------------------------------------------------------
/slora/server/router/req_queue.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/req_queue.py
--------------------------------------------------------------------------------
/slora/server/router/stats.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/stats.py
--------------------------------------------------------------------------------
/slora/server/router/vtc_req_queue.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/router/vtc_req_queue.py
--------------------------------------------------------------------------------
/slora/server/sampling_params.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/sampling_params.py
--------------------------------------------------------------------------------
/slora/server/tokenizer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/server/tokenizer.py
--------------------------------------------------------------------------------
/slora/utils/__init__.py:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/slora/utils/infer_utils.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/utils/infer_utils.py
--------------------------------------------------------------------------------
/slora/utils/metric.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/utils/metric.py
--------------------------------------------------------------------------------
/slora/utils/model_load.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/utils/model_load.py
--------------------------------------------------------------------------------
/slora/utils/model_utils.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/utils/model_utils.py
--------------------------------------------------------------------------------
/slora/utils/net_utils.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/slora/utils/net_utils.py
--------------------------------------------------------------------------------
/test/kernel/test_kernel_correctness.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/kernel/test_kernel_correctness.py
--------------------------------------------------------------------------------
/test/kernel/test_kernel_correctness_multi_rank.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/kernel/test_kernel_correctness_multi_rank.py
--------------------------------------------------------------------------------
/test/model/model_infer.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/model/model_infer.py
--------------------------------------------------------------------------------
/test/model/model_infer_multimodal.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/model/model_infer_multimodal.py
--------------------------------------------------------------------------------
/test/model/test_llama.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/model/test_llama.py
--------------------------------------------------------------------------------
/test/model/test_llama2.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/model/test_llama2.py
--------------------------------------------------------------------------------
/test/test_e2e/exp_suite.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/test_e2e/exp_suite.py
--------------------------------------------------------------------------------
/test/test_e2e/launch_server.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/test_e2e/launch_server.py
--------------------------------------------------------------------------------
/test/test_e2e/run_exp.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/test_e2e/run_exp.py
--------------------------------------------------------------------------------
/test/test_e2e/trace.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/S-LoRA/S-LoRA/HEAD/test/test_e2e/trace.py
--------------------------------------------------------------------------------