Bram Vanroy
commited on
Commit
·
d20fb33
1
Parent(s):
3b045b8
add evals
Browse files
evals/arc/arc_nl_Yi-6B.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"arc_nl": {
|
| 4 |
+
"acc": 0.2711719418306245,
|
| 5 |
+
"acc_stderr": 0.01300808888870032,
|
| 6 |
+
"acc_norm": 0.2951240376390077,
|
| 7 |
+
"acc_norm_stderr": 0.013345572865502647
|
| 8 |
+
}
|
| 9 |
+
},
|
| 10 |
+
"versions": {
|
| 11 |
+
"arc_nl": 0
|
| 12 |
+
},
|
| 13 |
+
"config": {
|
| 14 |
+
"model": "hf-auto",
|
| 15 |
+
"model_args": "pretrained=01-ai/Yi-6B,use_accelerate=True,device_map_option=auto,dtype=auto,load_in_8bit=True",
|
| 16 |
+
"batch_size": 2,
|
| 17 |
+
"device": "cuda",
|
| 18 |
+
"no_cache": false,
|
| 19 |
+
"limit": null,
|
| 20 |
+
"bootstrap_iters": 100000,
|
| 21 |
+
"description_dict": {}
|
| 22 |
+
}
|
| 23 |
+
}
|
evals/hellaswag/hellaswag_nl_Yi-6B.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"hellaswag_nl": {
|
| 4 |
+
"acc": 0.34225580140313006,
|
| 5 |
+
"acc_stderr": 0.004929520183215306,
|
| 6 |
+
"acc_norm": 0.41640582838640045,
|
| 7 |
+
"acc_norm_stderr": 0.005121705462751119
|
| 8 |
+
}
|
| 9 |
+
},
|
| 10 |
+
"versions": {
|
| 11 |
+
"hellaswag_nl": 1
|
| 12 |
+
},
|
| 13 |
+
"config": {
|
| 14 |
+
"model": "hf-auto",
|
| 15 |
+
"model_args": "pretrained=01-ai/Yi-6B,use_accelerate=True,device_map_option=auto,dtype=auto,load_in_8bit=True",
|
| 16 |
+
"batch_size": 2,
|
| 17 |
+
"device": "cuda",
|
| 18 |
+
"no_cache": false,
|
| 19 |
+
"limit": null,
|
| 20 |
+
"bootstrap_iters": 100000,
|
| 21 |
+
"description_dict": {}
|
| 22 |
+
}
|
| 23 |
+
}
|
evals/hellaswag/hellaswag_nl_phi-2.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"hellaswag_nl": {
|
| 4 |
+
"acc": 0.2746896923907178,
|
| 5 |
+
"acc_stderr": 0.004637499043161815,
|
| 6 |
+
"acc_norm": 0.3019967620075553,
|
| 7 |
+
"acc_norm_stderr": 0.004770134104673541
|
| 8 |
+
}
|
| 9 |
+
},
|
| 10 |
+
"versions": {
|
| 11 |
+
"hellaswag_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 |
+
}
|
evals/mmlu/mmlu_nl_Yi-6B-Chat.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"mmlu_nl": {
|
| 4 |
+
"acc": 0.4071488199134856,
|
| 5 |
+
"acc_stderr": 0.004280136715848525,
|
| 6 |
+
"acc_norm": 0.3481065492904303,
|
| 7 |
+
"acc_norm_stderr": 0.004150043156295438
|
| 8 |
+
}
|
| 9 |
+
},
|
| 10 |
+
"versions": {
|
| 11 |
+
"mmlu_nl": 0
|
| 12 |
+
},
|
| 13 |
+
"config": {
|
| 14 |
+
"model": "hf-auto",
|
| 15 |
+
"model_args": "pretrained=01-ai/Yi-6B-Chat,use_accelerate=True,device_map_option=auto,dtype=auto,load_in_8bit=True",
|
| 16 |
+
"batch_size": 2,
|
| 17 |
+
"device": "cuda",
|
| 18 |
+
"no_cache": false,
|
| 19 |
+
"limit": null,
|
| 20 |
+
"bootstrap_iters": 100000,
|
| 21 |
+
"description_dict": {}
|
| 22 |
+
}
|
| 23 |
+
}
|
evals/models.json
CHANGED
|
@@ -135,6 +135,22 @@
|
|
| 135 |
"num_parameters": 6738440192,
|
| 136 |
"quantization": "8-bit"
|
| 137 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 138 |
"yi-6b-chat": {
|
| 139 |
"compute_dtype": "auto",
|
| 140 |
"dutch_coverage": "none",
|
|
|
|
| 135 |
"num_parameters": 6738440192,
|
| 136 |
"quantization": "8-bit"
|
| 137 |
},
|
| 138 |
+
"phi-2": {
|
| 139 |
+
"compute_dtype": "auto",
|
| 140 |
+
"dutch_coverage": "none",
|
| 141 |
+
"model_name": "microsoft/phi-2",
|
| 142 |
+
"model_type": "instruction-tuned",
|
| 143 |
+
"num_parameters": 2779683840,
|
| 144 |
+
"quantization": "8-bit"
|
| 145 |
+
},
|
| 146 |
+
"yi-6b": {
|
| 147 |
+
"compute_dtype": "auto",
|
| 148 |
+
"dutch_coverage": "none",
|
| 149 |
+
"model_name": "01-ai/Yi-6B",
|
| 150 |
+
"model_type": "pretrained",
|
| 151 |
+
"num_parameters": 6061035520,
|
| 152 |
+
"quantization": "8-bit"
|
| 153 |
+
},
|
| 154 |
"yi-6b-chat": {
|
| 155 |
"compute_dtype": "auto",
|
| 156 |
"dutch_coverage": "none",
|
evals/truthfulqa/truthfulqa_nl_Yi-6B.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"truthfulqa_nl": {
|
| 4 |
+
"mc1": 0.2624203821656051,
|
| 5 |
+
"mc1_stderr": 0.015712496862048314,
|
| 6 |
+
"mc2": 0.4268865025462531,
|
| 7 |
+
"mc2_stderr": 0.015393899192344402
|
| 8 |
+
}
|
| 9 |
+
},
|
| 10 |
+
"versions": {
|
| 11 |
+
"truthfulqa_nl": 1
|
| 12 |
+
},
|
| 13 |
+
"config": {
|
| 14 |
+
"model": "hf-auto",
|
| 15 |
+
"model_args": "pretrained=01-ai/Yi-6B,use_accelerate=True,device_map_option=auto,dtype=auto,load_in_8bit=True",
|
| 16 |
+
"batch_size": 2,
|
| 17 |
+
"device": "cuda",
|
| 18 |
+
"no_cache": false,
|
| 19 |
+
"limit": null,
|
| 20 |
+
"bootstrap_iters": 100000,
|
| 21 |
+
"description_dict": {}
|
| 22 |
+
}
|
| 23 |
+
}
|