Update to 14-category MMLU-Pro model (77.82% accuracy) - fixes #584 #591
Browse files- category_mapping.json +34 -0
- config.json +29 -7
- label_mapping.json +28 -6
- model.safetensors +2 -2
category_mapping.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"category_to_idx": {
|
| 3 |
+
"biology": 0,
|
| 4 |
+
"business": 1,
|
| 5 |
+
"chemistry": 2,
|
| 6 |
+
"computer science": 3,
|
| 7 |
+
"economics": 4,
|
| 8 |
+
"engineering": 5,
|
| 9 |
+
"health": 6,
|
| 10 |
+
"history": 7,
|
| 11 |
+
"law": 8,
|
| 12 |
+
"math": 9,
|
| 13 |
+
"other": 10,
|
| 14 |
+
"philosophy": 11,
|
| 15 |
+
"physics": 12,
|
| 16 |
+
"psychology": 13
|
| 17 |
+
},
|
| 18 |
+
"idx_to_category": {
|
| 19 |
+
"0": "biology",
|
| 20 |
+
"1": "business",
|
| 21 |
+
"2": "chemistry",
|
| 22 |
+
"3": "computer science",
|
| 23 |
+
"4": "economics",
|
| 24 |
+
"5": "engineering",
|
| 25 |
+
"6": "health",
|
| 26 |
+
"7": "history",
|
| 27 |
+
"8": "law",
|
| 28 |
+
"9": "math",
|
| 29 |
+
"10": "other",
|
| 30 |
+
"11": "philosophy",
|
| 31 |
+
"12": "physics",
|
| 32 |
+
"13": "psychology"
|
| 33 |
+
}
|
| 34 |
+
}
|
config.json
CHANGED
|
@@ -10,16 +10,38 @@
|
|
| 10 |
"hidden_dropout_prob": 0.1,
|
| 11 |
"hidden_size": 768,
|
| 12 |
"id2label": {
|
| 13 |
-
"0": "
|
| 14 |
-
"1": "
|
| 15 |
-
"2": "
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
},
|
| 17 |
"initializer_range": 0.02,
|
| 18 |
"intermediate_size": 3072,
|
| 19 |
"label2id": {
|
| 20 |
-
"
|
| 21 |
-
"
|
| 22 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
},
|
| 24 |
"layer_norm_eps": 1e-12,
|
| 25 |
"max_position_embeddings": 512,
|
|
@@ -28,7 +50,7 @@
|
|
| 28 |
"num_hidden_layers": 12,
|
| 29 |
"pad_token_id": 0,
|
| 30 |
"position_embedding_type": "absolute",
|
| 31 |
-
"transformers_version": "4.
|
| 32 |
"type_vocab_size": 2,
|
| 33 |
"use_cache": true,
|
| 34 |
"vocab_size": 30522
|
|
|
|
| 10 |
"hidden_dropout_prob": 0.1,
|
| 11 |
"hidden_size": 768,
|
| 12 |
"id2label": {
|
| 13 |
+
"0": "biology",
|
| 14 |
+
"1": "business",
|
| 15 |
+
"2": "chemistry",
|
| 16 |
+
"3": "computer science",
|
| 17 |
+
"4": "economics",
|
| 18 |
+
"5": "engineering",
|
| 19 |
+
"6": "health",
|
| 20 |
+
"7": "history",
|
| 21 |
+
"8": "law",
|
| 22 |
+
"9": "math",
|
| 23 |
+
"10": "other",
|
| 24 |
+
"11": "philosophy",
|
| 25 |
+
"12": "physics",
|
| 26 |
+
"13": "psychology"
|
| 27 |
},
|
| 28 |
"initializer_range": 0.02,
|
| 29 |
"intermediate_size": 3072,
|
| 30 |
"label2id": {
|
| 31 |
+
"biology": 0,
|
| 32 |
+
"business": 1,
|
| 33 |
+
"chemistry": 2,
|
| 34 |
+
"computer science": 3,
|
| 35 |
+
"economics": 4,
|
| 36 |
+
"engineering": 5,
|
| 37 |
+
"health": 6,
|
| 38 |
+
"history": 7,
|
| 39 |
+
"law": 8,
|
| 40 |
+
"math": 9,
|
| 41 |
+
"other": 10,
|
| 42 |
+
"philosophy": 11,
|
| 43 |
+
"physics": 12,
|
| 44 |
+
"psychology": 13
|
| 45 |
},
|
| 46 |
"layer_norm_eps": 1e-12,
|
| 47 |
"max_position_embeddings": 512,
|
|
|
|
| 50 |
"num_hidden_layers": 12,
|
| 51 |
"pad_token_id": 0,
|
| 52 |
"position_embedding_type": "absolute",
|
| 53 |
+
"transformers_version": "4.57.1",
|
| 54 |
"type_vocab_size": 2,
|
| 55 |
"use_cache": true,
|
| 56 |
"vocab_size": 30522
|
label_mapping.json
CHANGED
|
@@ -1,12 +1,34 @@
|
|
| 1 |
{
|
| 2 |
"category_to_idx": {
|
| 3 |
-
"
|
| 4 |
-
"
|
| 5 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
},
|
| 7 |
"idx_to_category": {
|
| 8 |
-
"0": "
|
| 9 |
-
"1": "
|
| 10 |
-
"2": "
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
}
|
| 12 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"category_to_idx": {
|
| 3 |
+
"biology": 0,
|
| 4 |
+
"business": 1,
|
| 5 |
+
"chemistry": 2,
|
| 6 |
+
"computer science": 3,
|
| 7 |
+
"economics": 4,
|
| 8 |
+
"engineering": 5,
|
| 9 |
+
"health": 6,
|
| 10 |
+
"history": 7,
|
| 11 |
+
"law": 8,
|
| 12 |
+
"math": 9,
|
| 13 |
+
"other": 10,
|
| 14 |
+
"philosophy": 11,
|
| 15 |
+
"physics": 12,
|
| 16 |
+
"psychology": 13
|
| 17 |
},
|
| 18 |
"idx_to_category": {
|
| 19 |
+
"0": "biology",
|
| 20 |
+
"1": "business",
|
| 21 |
+
"2": "chemistry",
|
| 22 |
+
"3": "computer science",
|
| 23 |
+
"4": "economics",
|
| 24 |
+
"5": "engineering",
|
| 25 |
+
"6": "health",
|
| 26 |
+
"7": "history",
|
| 27 |
+
"8": "law",
|
| 28 |
+
"9": "math",
|
| 29 |
+
"10": "other",
|
| 30 |
+
"11": "philosophy",
|
| 31 |
+
"12": "physics",
|
| 32 |
+
"13": "psychology"
|
| 33 |
}
|
| 34 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7286cba260e36c42a02d41798841361f0c4b09f2fba8bc391dd3248e9ed11cdb
|
| 3 |
+
size 437995560
|