Update README.md

Former-commit-id: 2d43b8bb49057e14a9f79146acdcc0cfa94bcc5a
This commit is contained in:
hiyouga 2024-06-13 16:02:21 +08:00
parent f81a839197
commit d4ce280fbc

View File

@ -97,25 +97,25 @@ FORCE_TORCHRUN=1 llamafactory-cli train examples/train_lora/llama3_lora_sft_ds3.
#### Supervised Fine-Tuning with 4/8-bit Bitsandbytes Quantization (Recommended)
```bash
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train examples/train_qlora/llama3_lora_sft_bitsandbytes.yaml
llamafactory-cli train examples/train_qlora/llama3_lora_sft_bitsandbytes.yaml
```
#### Supervised Fine-Tuning with 4/8-bit GPTQ Quantization
```bash
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train examples/train_qlora/llama3_lora_sft_gptq.yaml
llamafactory-cli train examples/train_qlora/llama3_lora_sft_gptq.yaml
```
#### Supervised Fine-Tuning with 4-bit AWQ Quantization
```bash
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train examples/train_qlora/llama3_lora_sft_awq.yaml
llamafactory-cli train examples/train_qlora/llama3_lora_sft_awq.yaml
```
#### Supervised Fine-Tuning with 2-bit AQLM Quantization
```bash
CUDA_VISIBLE_DEVICES=0 llamafactory-cli train examples/train_qlora/llama3_lora_sft_aqlm.yaml
llamafactory-cli train examples/train_qlora/llama3_lora_sft_aqlm.yaml
```
### Full-Parameter Fine-Tuning