Hugging Face's logo

Akira1101
/
lora-structeval-Ver14

Text Generation
PEFT
Safetensors
English
qlora
lora
structured-output
Model card Files
xet
 Community
Akira1101 commited on
Commit
5ec47bf
·
verified ·
1 Parent(s): 8dc7b19

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -1
README.md CHANGED
@@ -37,6 +37,8 @@ while intermediate reasoning (Chain-of-Thought) is masked.
37
  - Epochs: 1
38
  - Learning rate: 5e-05
39
  - LoRA: r=2, alpha=2
 
 
40
 
41
  ## Usage
42
 
@@ -46,7 +48,7 @@ from peft import PeftModel
46
  import torch
47
 
48
  base = "Qwen/Qwen3-4B-Instruct-2507"
49
- adapter = "your_id/your-repo"
50
 
51
  tokenizer = AutoTokenizer.from_pretrained(base)
52
  model = AutoModelForCausalLM.from_pretrained(
 
37
  - Epochs: 1
38
  - Learning rate: 5e-05
39
  - LoRA: r=2, alpha=2
40
+ - SFT_WARMUP_RATIO = 0.5
41
+ - SFT WEIGHT_DECAY = 0.1
42
 
43
  ## Usage
44
 
 
48
  import torch
49
 
50
  base = "Qwen/Qwen3-4B-Instruct-2507"
51
+ adapter = "Akira1101/lora-structeval-Ver14"
52
 
53
  tokenizer = AutoTokenizer.from_pretrained(base)
54
  model = AutoModelForCausalLM.from_pretrained(