p04798526
/
LLaMA-Factory-310P3
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

update examples

This commit is contained in:
hiyouga 2024-06-27 00:53:33 +08:00
parent f17c9dfd84
commit d417e63f92
4 changed files with 6 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
examples/README.md
View File

@ -94,10 +94,10 @@ FORCE_TORCHRUN=1 llamafactory-cli train examples/train_lora/llama3_lora_sft_ds3.
### QLoRA Fine-Tuning ### QLoRA Fine-Tuning
#### Supervised Fine-Tuning with 4/8-bit Bitsandbytes Quantization (Recommended) #### Supervised Fine-Tuning with 4/8-bit Bitsandbytes/HQQ/EETQ Quantization (Recommended)
```bash ```bash
llamafactory-cli train examples/train_qlora/llama3_lora_sft_bitsandbytes.yaml llamafactory-cli train examples/train_qlora/llama3_lora_sft_otfq.yaml
``` ```
#### Supervised Fine-Tuning with 4/8-bit GPTQ Quantization #### Supervised Fine-Tuning with 4/8-bit GPTQ Quantization

4
examples/README_zh.md
View File

@ -94,10 +94,10 @@ FORCE_TORCHRUN=1 llamafactory-cli train examples/train_lora/llama3_lora_sft_ds3.
### QLoRA 微调 ### QLoRA 微调
#### 基于 4/8 比特 Bitsandbytes 量化进行指令监督微调（推荐） #### 基于 4/8 比特 Bitsandbytes/HQQ/EETQ 量化进行指令监督微调（推荐）
```bash ```bash
llamafactory-cli train examples/train_qlora/llama3_lora_sft_bitsandbytes.yaml llamafactory-cli train examples/train_qlora/llama3_lora_sft_otfq.yaml
``` ```
#### 基于 4/8 比特 GPTQ 量化进行指令监督微调 #### 基于 4/8 比特 GPTQ 量化进行指令监督微调

2
examples/train_lora/llama3_lora_dpo.yaml
View File

@ -7,7 +7,7 @@ do_train: true
finetuning_type: lora finetuning_type: lora
lora_target: all lora_target: all
pref_beta: 0.1 pref_beta: 0.1
pref_loss: sigmoid # [sigmoid (dpo), orpo, simpo] pref_loss: sigmoid # choices: [sigmoid (dpo), orpo, simpo]
### dataset ### dataset
dataset: dpo_en_demo dataset: dpo_en_demo

1
examples/train_qlora/llama3_lora_sft_bitsandbytes.yaml → examples/train_qlora/llama3_lora_sft_otfq.yaml
View File

@ -1,6 +1,7 @@
### model ### model
model_name_or_path: meta-llama/Meta-Llama-3-8B-Instruct model_name_or_path: meta-llama/Meta-Llama-3-8B-Instruct
quantization_bit: 4 quantization_bit: 4
quantization_method: bitsandbytes # choices: [bitsandbytes (4/8), hqq (2/3/4/5/6/8), eetq (8)]
### method ### method
stage: sft stage: sft