aiqwe commited on
Commit
fb5c65f
โ€ข
1 Parent(s): 61387c8

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -2
README.md CHANGED
@@ -58,14 +58,12 @@ for o in outputs:
58
  | Contents | Spec |
59
  |--------------------------------|-------------------------------------|
60
  | Base model | Qwen2.5-7B-Instruct |
61
- | Machine | A100 SXM 80GB ร— 2 |
62
  | dtype | bfloat16 |
63
  | PEFT | LoRA (r=8, alpha=64) |
64
  | Learning Rate | 1e-5 (varies by further training) |
65
  | LRScheduler | Cosine (warm-up: 0.05%) |
66
  | Optimizer | AdamW |
67
  | Distributed / Efficient Tuning | DeepSpeed v3, Flash Attention |
68
- | Global Batch Size | 128 |
69
 
70
  # Datset Card
71
  Reference ๋ฐ์ดํ„ฐ์…‹์€ ์ผ๋ถ€ ์ €์ž‘๊ถŒ ๊ด€๊ณ„๋กœ ์ธํ•ด Link๋กœ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.
 
58
  | Contents | Spec |
59
  |--------------------------------|-------------------------------------|
60
  | Base model | Qwen2.5-7B-Instruct |
 
61
  | dtype | bfloat16 |
62
  | PEFT | LoRA (r=8, alpha=64) |
63
  | Learning Rate | 1e-5 (varies by further training) |
64
  | LRScheduler | Cosine (warm-up: 0.05%) |
65
  | Optimizer | AdamW |
66
  | Distributed / Efficient Tuning | DeepSpeed v3, Flash Attention |
 
67
 
68
  # Datset Card
69
  Reference ๋ฐ์ดํ„ฐ์…‹์€ ์ผ๋ถ€ ์ €์ž‘๊ถŒ ๊ด€๊ณ„๋กœ ์ธํ•ด Link๋กœ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.