Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
vllm (pretrained=/root/autodl-tmp/Mistral-Nemo-Instruct-2407,add_bos_token=true,tensor_parallel_size=2,max_model_len=4096,gpu_memory_utilization=0.85,swap_space=0), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto
|
| 2 |
+
|Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
|
| 3 |
+
|-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
|
| 4 |
+
|gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.800|± |0.0253|
|
| 5 |
+
| | |strict-match | 5|exact_match|↑ |0.784|± |0.0261|
|
| 6 |
+
|
| 7 |
+
|
| 8 |
+
vllm (pretrained=/root/autodl-tmp/output,add_bos_token=true,tensor_parallel_size=2,max_model_len=2048), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto
|
| 9 |
+
|Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
|
| 10 |
+
|-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
|
| 11 |
+
|gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.792|± |0.0257|
|
| 12 |
+
| | |strict-match | 5|exact_match|↑ |0.776|± |0.0264|
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
I found some rules about quantization parameters and achieved better results.
|