Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model:
|
| 3 |
+
- gghfez/Mistral-Small-3.2-24B-Instruct-hf
|
| 4 |
+
---
|
| 5 |
+
vllm (pretrained=/root/autodl-tmp/Mistral-Small-3.2-24B-Instruct-hf,add_bos_token=true,max_model_len=3096,dtype=bfloat16,trust_remote_code=true), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto
|
| 6 |
+
|Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
|
| 7 |
+
|-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
|
| 8 |
+
|gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.908|± |0.0183|
|
| 9 |
+
| | |strict-match | 5|exact_match|↑ |0.904|± |0.0187|
|
| 10 |
+
|
| 11 |
+
vllm (pretrained=/root/autodl-tmp/Mistral-Small-3.2-24B-Instruct-hf,add_bos_token=true,max_model_len=3096,dtype=bfloat16,trust_remote_code=true,gpu_memory_utilization=0.8), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto
|
| 12 |
+
|Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
|
| 13 |
+
|-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
|
| 14 |
+
|gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.908|± |0.0129|
|
| 15 |
+
| | |strict-match | 5|exact_match|↑ |0.902|± |0.0133|
|
| 16 |
+
|
| 17 |
+
vllm (pretrained=/root/autodl-tmp/Mistral-Small-3.2-24B-Instruct-hf,add_bos_token=true,max_model_len=3048,dtype=bfloat16,trust_remote_code=true,gpu_memory_utilization=0.9), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: auto
|
| 18 |
+
| Groups |Version|Filter|n-shot|Metric| |Value | |Stderr|
|
| 19 |
+
|------------------|------:|------|------|------|---|-----:|---|-----:|
|
| 20 |
+
|mmlu | 2|none | |acc |↑ |0.8035|± |0.0129|
|
| 21 |
+
| - humanities | 2|none | |acc |↑ |0.8462|± |0.0247|
|
| 22 |
+
| - other | 2|none | |acc |↑ |0.8256|± |0.0262|
|
| 23 |
+
| - social sciences| 2|none | |acc |↑ |0.8389|± |0.0271|
|
| 24 |
+
| - stem | 2|none | |acc |↑ |0.7368|± |0.0246|
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
vllm (pretrained=/root/autodl-tmp/root90-128-4096-9.9999,add_bos_token=true,max_model_len=3096,dtype=bfloat16,trust_remote_code=true), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto
|
| 28 |
+
|Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
|
| 29 |
+
|-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
|
| 30 |
+
|gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.900|± |0.0190|
|
| 31 |
+
| | |strict-match | 5|exact_match|↑ |0.896|± |0.0193|
|
| 32 |
+
|
| 33 |
+
vllm (pretrained=/root/autodl-tmp/root90-128-4096-9.9999,add_bos_token=true,max_model_len=3096,dtype=bfloat16,trust_remote_code=true,gpu_memory_utilization=0.5), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto
|
| 34 |
+
|Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
|
| 35 |
+
|-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
|
| 36 |
+
|gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.892|± |0.0139|
|
| 37 |
+
| | |strict-match | 5|exact_match|↑ |0.886|± |0.0142|
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
vllm (pretrained=/root/autodl-tmp/root90-256-4096-9.9999,add_bos_token=true,max_model_len=3096,dtype=bfloat16,trust_remote_code=true,gpu_memory_utilization=0.5), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto
|
| 42 |
+
|Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
|
| 43 |
+
|-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
|
| 44 |
+
|gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.916|± |0.0176|
|
| 45 |
+
| | |strict-match | 5|exact_match|↑ |0.908|± |0.0183|
|
| 46 |
+
|
| 47 |
+
vllm (pretrained=/root/autodl-tmp/root90-256-4096-9.9999,add_bos_token=true,max_model_len=3096,dtype=bfloat16,trust_remote_code=true,gpu_memory_utilization=0.5), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto
|
| 48 |
+
|Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
|
| 49 |
+
|-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
|
| 50 |
+
|gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.904|± |0.0132|
|
| 51 |
+
| | |strict-match | 5|exact_match|↑ |0.898|± |0.0135|
|
| 52 |
+
|
| 53 |
+
vllm (pretrained=/root/autodl-tmp/root90-256-4096-9.9999,add_bos_token=true,max_model_len=3048,dtype=bfloat16,trust_remote_code=true,gpu_memory_utilization=0.9), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: auto
|
| 54 |
+
| Groups |Version|Filter|n-shot|Metric| |Value | |Stderr|
|
| 55 |
+
|------------------|------:|------|------|------|---|-----:|---|-----:|
|
| 56 |
+
|mmlu | 2|none | |acc |↑ |0.7895|± |0.0132|
|
| 57 |
+
| - humanities | 2|none | |acc |↑ |0.8256|± |0.0251|
|
| 58 |
+
| - other | 2|none | |acc |↑ |0.8051|± |0.0273|
|
| 59 |
+
| - social sciences| 2|none | |acc |↑ |0.7889|± |0.0292|
|
| 60 |
+
| - stem | 2|none | |acc |↑ |0.7544|± |0.0241|
|