├── README.md ├── agents ├── README.md ├── chatGPT_memory_agent.py ├── gemini_memory_agent.py ├── gemini_personalized_agent.py ├── personalized_agent.py ├── search_agent.py └── search_agent_gemini.py ├── pretrain ├── 124M_10B_log │ └── log.txt ├── README.md ├── fineweb.py ├── hellaswag.py ├── log_12_12_12_6e-4 │ └── log.txt ├── log_12_12_4_6e-4 │ └── log.txt ├── log_12_12_6_6e-4 │ └── log.txt ├── log_16_12_4_6e-4 │ └── log.txt ├── log_16_12_6_6e-4 │ └── log.txt ├── log_24_12_6_6e-04 │ └── log.txt ├── log_24_16_8_6e-04 │ └── log.txt ├── log_mix_16_layer(10)_attenH(8)_kvH(8)_acoef(0.02)_b(32)_gradacc(2)_step(38146)dropout(0.0)_jitter(0.0)exp(8)_exppt(2)_lr3e-03 │ └── log.txt ├── log_mix_16_layer(12)_attenH(12)_kvH(4)_acoef(0.02)_b(32)_gradacc(1)_step(38146)dropout(0.0)_jitter(0.0)exp(8)_exppt(2)_lr3e-03 │ └── log.txt ├── mixtral_result.ipynb ├── play_with_llama.ipynb ├── train_deepseek.py ├── train_gpt2.py ├── train_llama.py ├── train_llama_fsdp.py └── train_mixtral.py ├── rl ├── dpo.py └── dpo_demo.ipynb └── sft ├── LICENSE ├── README.md ├── axolotl_configs └── llama7b_tangshi.yml ├── configs ├── chat_format_test.yml ├── deepspeed.yaml ├── ds_config_stage1.json ├── inference_mixtral.yml ├── llama2_tangshi.yml ├── merge_mixtral_7b.yml ├── mistral-dpo.yml ├── mixtral_7b_tangshi.yml ├── mixtral_7b_tangshi_chatlm.yml ├── multiple_gpu.yaml ├── phi2-dolly.yml ├── yi-6b-dpo.yml ├── yi-6b-large.yml ├── yi-6b.yml ├── zero_stage1_config.json ├── zero_stage2_config.json └── zero_stage3_config.json ├── dataloader.py ├── inference-quantize.ipynb ├── inference.py ├── inference_qlora_finetune.ipynb ├── lc.py ├── learn_nanogpt.ipynb ├── learn_qlora_finetune.ipynb ├── merge_lora.py ├── mistral ├── build_data.py ├── mis-result.ipynb ├── mistral-ft.ipynb ├── test_merged.ipynb ├── train.csv └── val.csv ├── quantize.ipynb ├── requirements.txt ├── trl_finetune.py └── utils.py /README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/README.md -------------------------------------------------------------------------------- /agents/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/agents/README.md -------------------------------------------------------------------------------- /agents/chatGPT_memory_agent.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/agents/chatGPT_memory_agent.py -------------------------------------------------------------------------------- /agents/gemini_memory_agent.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/agents/gemini_memory_agent.py -------------------------------------------------------------------------------- /agents/gemini_personalized_agent.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/agents/gemini_personalized_agent.py -------------------------------------------------------------------------------- /agents/personalized_agent.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/agents/personalized_agent.py -------------------------------------------------------------------------------- /agents/search_agent.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/agents/search_agent.py -------------------------------------------------------------------------------- /agents/search_agent_gemini.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/agents/search_agent_gemini.py -------------------------------------------------------------------------------- /pretrain/124M_10B_log/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/124M_10B_log/log.txt -------------------------------------------------------------------------------- /pretrain/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/README.md -------------------------------------------------------------------------------- /pretrain/fineweb.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/fineweb.py -------------------------------------------------------------------------------- /pretrain/hellaswag.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/hellaswag.py -------------------------------------------------------------------------------- /pretrain/log_12_12_12_6e-4/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/log_12_12_12_6e-4/log.txt -------------------------------------------------------------------------------- /pretrain/log_12_12_4_6e-4/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/log_12_12_4_6e-4/log.txt -------------------------------------------------------------------------------- /pretrain/log_12_12_6_6e-4/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/log_12_12_6_6e-4/log.txt -------------------------------------------------------------------------------- /pretrain/log_16_12_4_6e-4/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/log_16_12_4_6e-4/log.txt -------------------------------------------------------------------------------- /pretrain/log_16_12_6_6e-4/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/log_16_12_6_6e-4/log.txt -------------------------------------------------------------------------------- /pretrain/log_24_12_6_6e-04/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/log_24_12_6_6e-04/log.txt -------------------------------------------------------------------------------- /pretrain/log_24_16_8_6e-04/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/log_24_16_8_6e-04/log.txt -------------------------------------------------------------------------------- /pretrain/log_mix_16_layer(10)_attenH(8)_kvH(8)_acoef(0.02)_b(32)_gradacc(2)_step(38146)dropout(0.0)_jitter(0.0)exp(8)_exppt(2)_lr3e-03/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/log_mix_16_layer(10)_attenH(8)_kvH(8)_acoef(0.02)_b(32)_gradacc(2)_step(38146)dropout(0.0)_jitter(0.0)exp(8)_exppt(2)_lr3e-03/log.txt -------------------------------------------------------------------------------- /pretrain/log_mix_16_layer(12)_attenH(12)_kvH(4)_acoef(0.02)_b(32)_gradacc(1)_step(38146)dropout(0.0)_jitter(0.0)exp(8)_exppt(2)_lr3e-03/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/log_mix_16_layer(12)_attenH(12)_kvH(4)_acoef(0.02)_b(32)_gradacc(1)_step(38146)dropout(0.0)_jitter(0.0)exp(8)_exppt(2)_lr3e-03/log.txt -------------------------------------------------------------------------------- /pretrain/mixtral_result.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/mixtral_result.ipynb -------------------------------------------------------------------------------- /pretrain/play_with_llama.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/play_with_llama.ipynb -------------------------------------------------------------------------------- /pretrain/train_deepseek.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/train_deepseek.py -------------------------------------------------------------------------------- /pretrain/train_gpt2.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/train_gpt2.py -------------------------------------------------------------------------------- /pretrain/train_llama.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/train_llama.py -------------------------------------------------------------------------------- /pretrain/train_llama_fsdp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/train_llama_fsdp.py -------------------------------------------------------------------------------- /pretrain/train_mixtral.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/pretrain/train_mixtral.py -------------------------------------------------------------------------------- /rl/dpo.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/rl/dpo.py -------------------------------------------------------------------------------- /rl/dpo_demo.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/rl/dpo_demo.ipynb -------------------------------------------------------------------------------- /sft/LICENSE: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/LICENSE -------------------------------------------------------------------------------- /sft/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/README.md -------------------------------------------------------------------------------- /sft/axolotl_configs/llama7b_tangshi.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/axolotl_configs/llama7b_tangshi.yml -------------------------------------------------------------------------------- /sft/configs/chat_format_test.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/chat_format_test.yml -------------------------------------------------------------------------------- /sft/configs/deepspeed.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/deepspeed.yaml -------------------------------------------------------------------------------- /sft/configs/ds_config_stage1.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/ds_config_stage1.json -------------------------------------------------------------------------------- /sft/configs/inference_mixtral.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/inference_mixtral.yml -------------------------------------------------------------------------------- /sft/configs/llama2_tangshi.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/llama2_tangshi.yml -------------------------------------------------------------------------------- /sft/configs/merge_mixtral_7b.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/merge_mixtral_7b.yml -------------------------------------------------------------------------------- /sft/configs/mistral-dpo.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/mistral-dpo.yml -------------------------------------------------------------------------------- /sft/configs/mixtral_7b_tangshi.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/mixtral_7b_tangshi.yml -------------------------------------------------------------------------------- /sft/configs/mixtral_7b_tangshi_chatlm.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/mixtral_7b_tangshi_chatlm.yml -------------------------------------------------------------------------------- /sft/configs/multiple_gpu.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/multiple_gpu.yaml -------------------------------------------------------------------------------- /sft/configs/phi2-dolly.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/phi2-dolly.yml -------------------------------------------------------------------------------- /sft/configs/yi-6b-dpo.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/yi-6b-dpo.yml -------------------------------------------------------------------------------- /sft/configs/yi-6b-large.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/yi-6b-large.yml -------------------------------------------------------------------------------- /sft/configs/yi-6b.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/yi-6b.yml -------------------------------------------------------------------------------- /sft/configs/zero_stage1_config.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/zero_stage1_config.json -------------------------------------------------------------------------------- /sft/configs/zero_stage2_config.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/zero_stage2_config.json -------------------------------------------------------------------------------- /sft/configs/zero_stage3_config.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/configs/zero_stage3_config.json -------------------------------------------------------------------------------- /sft/dataloader.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/dataloader.py -------------------------------------------------------------------------------- /sft/inference-quantize.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/inference-quantize.ipynb -------------------------------------------------------------------------------- /sft/inference.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/inference.py -------------------------------------------------------------------------------- /sft/inference_qlora_finetune.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/inference_qlora_finetune.ipynb -------------------------------------------------------------------------------- /sft/lc.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/lc.py -------------------------------------------------------------------------------- /sft/learn_nanogpt.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/learn_nanogpt.ipynb -------------------------------------------------------------------------------- /sft/learn_qlora_finetune.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/learn_qlora_finetune.ipynb -------------------------------------------------------------------------------- /sft/merge_lora.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/merge_lora.py -------------------------------------------------------------------------------- /sft/mistral/build_data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/mistral/build_data.py -------------------------------------------------------------------------------- /sft/mistral/mis-result.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/mistral/mis-result.ipynb -------------------------------------------------------------------------------- /sft/mistral/mistral-ft.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/mistral/mistral-ft.ipynb -------------------------------------------------------------------------------- /sft/mistral/test_merged.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/mistral/test_merged.ipynb -------------------------------------------------------------------------------- /sft/mistral/train.csv: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/mistral/train.csv -------------------------------------------------------------------------------- /sft/mistral/val.csv: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/mistral/val.csv -------------------------------------------------------------------------------- /sft/quantize.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/quantize.ipynb -------------------------------------------------------------------------------- /sft/requirements.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/requirements.txt -------------------------------------------------------------------------------- /sft/trl_finetune.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/trl_finetune.py -------------------------------------------------------------------------------- /sft/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hengjiUSTC/learn-llm/HEAD/sft/utils.py --------------------------------------------------------------------------------