Hugging Face's logo

noneUsername
/
Qwen2.5-Lumen-14B-W8A8-Dynamic-Per-Token

Safetensors
qwen2
8-bit precision
compressed-tensors
Model card Files Community
noneUsername commited on
Commit
2c27c51
·
verified ·
1 Parent(s): f8b26e9

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +16 -0
README.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model:
3
+ - v000000/Qwen2.5-Lumen-14B
4
+ ---
5
+
6
+ vllm (pretrained=/root/autodl-tmp/Qwen2.5-Lumen-14B,add_bos_token=true,tensor_parallel_size=2,max_model_len=2048,gpu_memory_utilization=0.99), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto
7
+ |Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
8
+ |-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
9
+ |gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.544|± |0.0316|
10
+ | | |strict-match | 5|exact_match|↑ |0.836|± |0.0235|
11
+
12
+ vllm (pretrained=/root/autodl-tmp/output,add_bos_token=true,tensor_parallel_size=2,max_model_len=2048,gpu_memory_utilization=0.99), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: 1
13
+ |Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
14
+ |-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
15
+ |gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.532|± |0.0316|
16
+ | | |strict-match | 5|exact_match|↑ |0.824|± |0.0241|