Bram Vanroy
commited on
Commit
·
2e653e4
1
Parent(s):
7579c54
complete Phi 2
Browse files- evals/arc/arc_nl_phi-2.json +23 -0
- evals/mmlu/mmlu_nl_phi-2.json +23 -0
- evals/models.json +1 -1
- evals/truthfulqa/truthfulqa_nl_phi-2.json +23 -0
evals/arc/arc_nl_phi-2.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"arc_nl": {
|
| 4 |
+
"acc": 0.20017108639863132,
|
| 5 |
+
"acc_stderr": 0.011707867575185044,
|
| 6 |
+
"acc_norm": 0.23781009409751924,
|
| 7 |
+
"acc_norm_stderr": 0.012457339987221747
|
| 8 |
+
}
|
| 9 |
+
},
|
| 10 |
+
"versions": {
|
| 11 |
+
"arc_nl": 0
|
| 12 |
+
},
|
| 13 |
+
"config": {
|
| 14 |
+
"model": "hf-auto",
|
| 15 |
+
"model_args": "pretrained=microsoft/phi-2,use_accelerate=True,device_map_option=auto,dtype=auto,load_in_8bit=True",
|
| 16 |
+
"batch_size": 4,
|
| 17 |
+
"device": "cuda",
|
| 18 |
+
"no_cache": false,
|
| 19 |
+
"limit": null,
|
| 20 |
+
"bootstrap_iters": 100000,
|
| 21 |
+
"description_dict": {}
|
| 22 |
+
}
|
| 23 |
+
}
|
evals/mmlu/mmlu_nl_phi-2.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"mmlu_nl": {
|
| 4 |
+
"acc": 0.2971085983152463,
|
| 5 |
+
"acc_stderr": 0.003981162755455281,
|
| 6 |
+
"acc_norm": 0.30006830082719893,
|
| 7 |
+
"acc_norm_stderr": 0.0039925107832952484
|
| 8 |
+
}
|
| 9 |
+
},
|
| 10 |
+
"versions": {
|
| 11 |
+
"mmlu_nl": 0
|
| 12 |
+
},
|
| 13 |
+
"config": {
|
| 14 |
+
"model": "hf-auto",
|
| 15 |
+
"model_args": "pretrained=microsoft/phi-2,use_accelerate=True,device_map_option=auto,dtype=auto,load_in_8bit=True",
|
| 16 |
+
"batch_size": 4,
|
| 17 |
+
"device": "cuda",
|
| 18 |
+
"no_cache": false,
|
| 19 |
+
"limit": null,
|
| 20 |
+
"bootstrap_iters": 100000,
|
| 21 |
+
"description_dict": {}
|
| 22 |
+
}
|
| 23 |
+
}
|
evals/models.json
CHANGED
|
@@ -167,4 +167,4 @@
|
|
| 167 |
"num_parameters": 7241732096,
|
| 168 |
"quantization": "8-bit"
|
| 169 |
}
|
| 170 |
-
}
|
|
|
|
| 167 |
"num_parameters": 7241732096,
|
| 168 |
"quantization": "8-bit"
|
| 169 |
}
|
| 170 |
+
}
|
evals/truthfulqa/truthfulqa_nl_phi-2.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"truthfulqa_nl": {
|
| 4 |
+
"mc1": 0.2229299363057325,
|
| 5 |
+
"mc1_stderr": 0.01486470359663736,
|
| 6 |
+
"mc2": 0.3977100255471215,
|
| 7 |
+
"mc2_stderr": 0.01601228004768883
|
| 8 |
+
}
|
| 9 |
+
},
|
| 10 |
+
"versions": {
|
| 11 |
+
"truthfulqa_nl": 1
|
| 12 |
+
},
|
| 13 |
+
"config": {
|
| 14 |
+
"model": "hf-auto",
|
| 15 |
+
"model_args": "pretrained=microsoft/phi-2,use_accelerate=True,device_map_option=auto,dtype=auto,load_in_8bit=True",
|
| 16 |
+
"batch_size": 4,
|
| 17 |
+
"device": "cuda",
|
| 18 |
+
"no_cache": false,
|
| 19 |
+
"limit": null,
|
| 20 |
+
"bootstrap_iters": 100000,
|
| 21 |
+
"description_dict": {}
|
| 22 |
+
}
|
| 23 |
+
}
|