ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

update webui and add CLIs

Browse Source 
Former-commit-id: 1368dda22ab875914c9dd86ee5146a4f6a4736ad
This commit is contained in:
hiyouga
2024-05-03 02:58:23 +08:00
parent 2cedb59bee
commit ce8200ad98
65 changed files with 363 additions and 372 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
examples/extras/badam/sft.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/extras/fsdp_qlora/sft.sh
View File

@@ -7,7 +7,7 @@ pip install "bitsandbytes>=0.43.0"
CUDA_VISIBLE_DEVICES=0,1 accelerate launch \
--config_file ../../accelerate/fsdp_config.yaml \
../../../src/train_bash.py \
../../../src/train.py \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-70b-hf \

2
examples/extras/galore/sft.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/extras/llama_pro/sft.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path ../../../models/llama2-7b-pro \

2
examples/extras/loraplus/sft.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/extras/mod/sft.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/full_multi_gpu/multi_node.sh
View File

@@ -6,7 +6,7 @@ python -m torch.distributed.run \
--node_rank $RANK \
--master_addr $MASTER_ADDR \
--master_port $MASTER_PORT \
../../src/train_bash.py \
../../src/train.py \
--deepspeed ../deepspeed/ds_z3_config.json \
--stage sft \
--do_train \

2
examples/full_multi_gpu/predict.sh
View File

@@ -2,7 +2,7 @@
CUDA_VISIBLE_DEVICES=0,1,2,3 accelerate launch \
--config_file ../accelerate/single_config.yaml \
../../src/train_bash.py \
../../src/train.py \
--stage sft \
--do_predict \
--model_name_or_path ../../saves/LLaMA2-7B/full/sft \

2
examples/full_multi_gpu/single_node.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
deepspeed --num_gpus 4 ../../src/train_bash.py \
deepspeed --num_gpus 4 ../../src/train.py \
--deepspeed ../deepspeed/ds_z3_config.json \
--stage sft \
--do_train \

2
examples/inference/api_demo.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 API_PORT=8000 python ../../src/api_demo.py \
CUDA_VISIBLE_DEVICES=0 API_PORT=8000 llamafactory-cli api \
--model_name_or_path meta-llama/Llama-2-7b-hf \
--adapter_name_or_path ../../saves/LLaMA2-7B/lora/sft \
--template default \

2
examples/inference/cli_demo.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/cli_demo.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli chat \
--model_name_or_path meta-llama/Llama-2-7b-hf \
--adapter_name_or_path ../../saves/LLaMA2-7B/lora/sft \
--template default \

2
examples/inference/evaluate.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/evaluate.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli eval \
--model_name_or_path meta-llama/Llama-2-7b-hf \
--adapter_name_or_path ../../saves/LLaMA2-7B/lora/sft \
--template fewshot \

2
examples/inference/web_demo.sh
View File

@@ -1,7 +1,7 @@
#!/bin/bash
# add `--visual_inputs True` to load MLLM
CUDA_VISIBLE_DEVICES=0 python ../../src/web_demo.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli webchat \
--model_name_or_path meta-llama/Llama-2-7b-hf \
--adapter_name_or_path ../../saves/LLaMA2-7B/lora/sft \
--template default \

3
examples/lora_multi_gpu/ds_zero3.sh
View File

@@ -1,6 +1,7 @@
#!/bin/bash
# ZeRO-3 enables weight sharding on multiple GPUs
deepspeed --num_gpus 4 ../../src/train_bash.py \
deepspeed --num_gpus 4 ../../src/train.py \
--deepspeed ../deepspeed/ds_z3_config.json \
--stage sft \
--do_train \

2
examples/lora_multi_gpu/multi_node.sh
View File

@@ -3,7 +3,7 @@
CUDA_VISIBLE_DEVICES=0,1,2,3 accelerate launch \
--config_file ../accelerate/master_config.yaml \
../../src/train_bash.py \
../../src/train.py \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_multi_gpu/single_node.sh
View File

@@ -2,7 +2,7 @@
CUDA_VISIBLE_DEVICES=0,1,2,3 accelerate launch \
--config_file ../accelerate/single_config.yaml \
../../src/train_bash.py \
../../src/train.py \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_single_gpu/dpo.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage dpo \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_single_gpu/orpo.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage orpo \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_single_gpu/ppo.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage ppo \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_single_gpu/predict.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_predict \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_single_gpu/prepare.sh
View File

@@ -1,7 +1,7 @@
#!/bin/bash
# use `--tokenized_path` in training script to load data
CUDA_VISIBLE_DEVICES= python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES= llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_single_gpu/pretrain.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage pt \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_single_gpu/reward.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage rm \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_single_gpu/sft.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/lora_single_gpu/sft_mllm.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path llava-hf/llava-1.5-7b-hf \

2
examples/merge_lora/merge.sh
View File

@@ -1,7 +1,7 @@
#!/bin/bash
# DO NOT use quantized model or quantization_bit when merging lora weights
CUDA_VISIBLE_DEVICES=0 python ../../src/export_model.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli export \
--model_name_or_path meta-llama/Llama-2-7b-hf \
--adapter_name_or_path ../../saves/LLaMA2-7B/lora/sft \
--template default \

2
examples/merge_lora/quantize.sh
View File

@@ -1,7 +1,7 @@
#!/bin/bash
# NEED TO run `merge.sh` before using this script
CUDA_VISIBLE_DEVICES=0 python ../../src/export_model.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli export \
--model_name_or_path ../../models/llama2-7b-sft \
--template default \
--export_dir ../../models/llama2-7b-sft-int4 \

2
examples/qlora_single_gpu/aqlm.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path BlackSamorez/Llama-2-7b-AQLM-2Bit-1x16-hf \

2
examples/qlora_single_gpu/awq.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path TheBloke/Llama-2-7B-AWQ \

2
examples/qlora_single_gpu/bitsandbytes.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path meta-llama/Llama-2-7b-hf \

2
examples/qlora_single_gpu/gptq.sh
View File

@@ -1,6 +1,6 @@
#!/bin/bash
CUDA_VISIBLE_DEVICES=0 python ../../src/train_bash.py \
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train \
--stage sft \
--do_train \
--model_name_or_path TheBloke/Llama-2-7B-GPTQ \