ACDRepo commited on
Commit
eb03713
·
verified ·
1 Parent(s): 9d72d96

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -1,35 +1,3 @@
1
- *.7z filter=lfs diff=lfs merge=lfs -text
2
- *.arrow filter=lfs diff=lfs merge=lfs -text
3
  *.bin filter=lfs diff=lfs merge=lfs -text
4
- *.bz2 filter=lfs diff=lfs merge=lfs -text
5
- *.ckpt filter=lfs diff=lfs merge=lfs -text
6
- *.ftz filter=lfs diff=lfs merge=lfs -text
7
- *.gz filter=lfs diff=lfs merge=lfs -text
8
- *.h5 filter=lfs diff=lfs merge=lfs -text
9
- *.joblib filter=lfs diff=lfs merge=lfs -text
10
- *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
- *.mlmodel filter=lfs diff=lfs merge=lfs -text
12
- *.model filter=lfs diff=lfs merge=lfs -text
13
- *.msgpack filter=lfs diff=lfs merge=lfs -text
14
- *.npy filter=lfs diff=lfs merge=lfs -text
15
- *.npz filter=lfs diff=lfs merge=lfs -text
16
- *.onnx filter=lfs diff=lfs merge=lfs -text
17
- *.ot filter=lfs diff=lfs merge=lfs -text
18
- *.parquet filter=lfs diff=lfs merge=lfs -text
19
- *.pb filter=lfs diff=lfs merge=lfs -text
20
- *.pickle filter=lfs diff=lfs merge=lfs -text
21
- *.pkl filter=lfs diff=lfs merge=lfs -text
22
  *.pt filter=lfs diff=lfs merge=lfs -text
23
- *.pth filter=lfs diff=lfs merge=lfs -text
24
- *.rar filter=lfs diff=lfs merge=lfs -text
25
- *.safetensors filter=lfs diff=lfs merge=lfs -text
26
- saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
- *.tar.* filter=lfs diff=lfs merge=lfs -text
28
- *.tar filter=lfs diff=lfs merge=lfs -text
29
- *.tflite filter=lfs diff=lfs merge=lfs -text
30
- *.tgz filter=lfs diff=lfs merge=lfs -text
31
- *.wasm filter=lfs diff=lfs merge=lfs -text
32
- *.xz filter=lfs diff=lfs merge=lfs -text
33
- *.zip filter=lfs diff=lfs merge=lfs -text
34
- *.zst filter=lfs diff=lfs merge=lfs -text
35
- *tfevents* filter=lfs diff=lfs merge=lfs -text
 
1
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
 
2
  *.bin filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  *.pt filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
.gitignore ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ training_state.pt
2
+ .DS_Store
README.md CHANGED
@@ -1,5 +1,5 @@
1
  ---
2
- license: cc-by-4.0
3
  pipeline_tag: text-generation
4
  tags:
5
  - math
@@ -36,11 +36,11 @@ PermuFormer was trained autoregressively on synthetic permutation examples gener
36
 
37
  Training tasks cover three broad families:
38
 
39
- - **Translation between encodings:** One-line notation, cycle notation, reduced Coxeter expressions, RSK tableaux, inversion vectors, and Lehmer codes.
40
- - **Permutation statistics and properties:** Length, descents, fixed points, sign/parity, cycle type, RSK shape, pattern avoidance, longest increasing/decreasing subsequences, and related statistics.
41
- - **Algebraic operations and comparisons:** Product/composition, inverse, powers, conjugation, commutator, relative products, multiplication by simple transpositions, complement, reverse, descent tests, and Bruhat order.
42
 
43
- Some targets include computational witnesses before the final answer, for example inversions are listed before a computation of length. Similarly pattern witnesses are stated before an avoidance answer.
44
 
45
  ## Usage
46
 
@@ -50,7 +50,7 @@ Use deterministic decoding for most evaluation-style tasks. Make sure special to
50
  from transformers import AutoModelForCausalLM, AutoTokenizer
51
  import torch
52
 
53
- model_id = "ACDRepo/permuformer"
54
 
55
  tokenizer = AutoTokenizer.from_pretrained(model_id)
56
  model = AutoModelForCausalLM.from_pretrained(model_id)
@@ -100,14 +100,16 @@ Algebraic operation example:
100
 
101
  ## Evaluation Notes
102
 
103
- The training code evaluates by exact match on the generated right-hand side after `=`.
104
 
105
  - Overall exact match: **98.44%**
106
  - Translation: **97.78%**
107
  - Property/statistic tasks: **99.17%**
108
  - Algebraic tasks: **98.36%**
109
 
110
- PermuFormer is substantially more accurate than frontier general-purpose LLMs on a small held-out sample from the model's symbolic test distribution (though note that this comparison is imperfect because PermuFormer is trained directly in this syntax and on these tasks).
 
 
111
 
112
  ## Finetuning
113
 
@@ -118,6 +120,8 @@ PermuFormer is designed to be finetuned on specialized permutation tasks. Experi
118
  - Schubert polynomial structure constants
119
  - Kazhdan-Lusztig polynomial degree prediction
120
 
 
 
121
  ## Limitations
122
 
123
  - This is a specialist symbolic model. It expects the exact whitespace-tokenized syntax used during training and is brittle to natural-language paraphrases or malformed prompts.
@@ -128,4 +132,4 @@ PermuFormer is designed to be finetuned on specialized permutation tasks. Experi
128
 
129
  ## Citation
130
 
131
- Coming soon.
 
1
  ---
2
+ library_name: transformers
3
  pipeline_tag: text-generation
4
  tags:
5
  - math
 
36
 
37
  Training tasks cover three broad families:
38
 
39
+ - **Translation between encodings:** one-line notation, cycle notation, reduced Coxeter expressions, RSK tableaux, inversion vectors, and Lehmer codes.
40
+ - **Permutation statistics and properties:** length, descents, fixed points, sign/parity, cycle type, RSK shape, pattern avoidance, longest increasing/decreasing subsequences, and related statistics.
41
+ - **Algebraic operations and comparisons:** product/composition, inverse, powers, conjugation, commutator, relative products, multiplication by simple transpositions, complement, reverse, descent tests, and Bruhat order.
42
 
43
+ Some targets include computational witnesses before the final answer, for example inversion lists before a length answer or pattern witnesses before an avoidance answer.
44
 
45
  ## Usage
46
 
 
50
  from transformers import AutoModelForCausalLM, AutoTokenizer
51
  import torch
52
 
53
+ model_id = "YOUR_ORG/permuformer"
54
 
55
  tokenizer = AutoTokenizer.from_pretrained(model_id)
56
  model = AutoModelForCausalLM.from_pretrained(model_id)
 
100
 
101
  ## Evaluation Notes
102
 
103
+ The training code evaluates by exact match on the generated right-hand side after `=`. The local training log for this repository reports, at step 2,522,000 on a 2,560-example stratified evaluation sample:
104
 
105
  - Overall exact match: **98.44%**
106
  - Translation: **97.78%**
107
  - Property/statistic tasks: **99.17%**
108
  - Algebraic tasks: **98.36%**
109
 
110
+ These figures are from the local log and should be treated as checkpoint-adjacent repository metadata, not a full benchmark report for every downstream setting.
111
+
112
+ The paper also reports that PermuFormer is substantially more accurate than frontier general-purpose LLMs on a small held-out sample from the model's symbolic test distribution, while noting that the comparison is imperfect because PermuFormer was trained directly in this syntax.
113
 
114
  ## Finetuning
115
 
 
120
  - Schubert polynomial structure constants
121
  - Kazhdan-Lusztig polynomial degree prediction
122
 
123
+ The repository's finetuning scripts compare starting from this pretrained checkpoint with training the same architecture from scratch.
124
+
125
  ## Limitations
126
 
127
  - This is a specialist symbolic model. It expects the exact whitespace-tokenized syntax used during training and is brittle to natural-language paraphrases or malformed prompts.
 
132
 
133
  ## Citation
134
 
135
+ If you use this model, please cite the accompanying PermuFormer paper once citation details are available.
config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 5,
8
+ "dtype": "float32",
9
+ "eos_token_id": 5,
10
+ "head_dim": 64,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 768,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 2048,
15
+ "max_position_embeddings": 2048,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "num_attention_heads": 12,
19
+ "num_hidden_layers": 12,
20
+ "num_key_value_heads": 4,
21
+ "pad_token_id": 185,
22
+ "pretraining_tp": 1,
23
+ "rms_norm_eps": 1e-05,
24
+ "rope_scaling": null,
25
+ "rope_theta": 10000.0,
26
+ "tie_word_embeddings": true,
27
+ "transformers_version": "4.57.0",
28
+ "use_cache": true,
29
+ "vocab_size": 186
30
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 5,
4
+ "eos_token_id": 5,
5
+ "pad_token_id": 185,
6
+ "transformers_version": "4.57.0"
7
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ad9bb41e00e129a91e2150ad67c1db289801fdb86b94e46d8f5c93516feb59f
3
+ size 302650224
special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<|endoftext|>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|endoftext|>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<pad>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "<unk>",
25
+ "lstrip": false,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ }
30
+ }
tokenizer.json ADDED
@@ -0,0 +1,244 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "1.0",
3
+ "truncation": {
4
+ "direction": "Right",
5
+ "max_length": 1000,
6
+ "strategy": "LongestFirst",
7
+ "stride": 0
8
+ },
9
+ "padding": {
10
+ "strategy": "BatchLongest",
11
+ "direction": "Right",
12
+ "pad_to_multiple_of": null,
13
+ "pad_id": 185,
14
+ "pad_type_id": 0,
15
+ "pad_token": "<pad>"
16
+ },
17
+ "added_tokens": [
18
+ {
19
+ "id": 5,
20
+ "content": "<|endoftext|>",
21
+ "single_word": false,
22
+ "lstrip": false,
23
+ "rstrip": false,
24
+ "normalized": false,
25
+ "special": true
26
+ },
27
+ {
28
+ "id": 184,
29
+ "content": "<unk>",
30
+ "single_word": false,
31
+ "lstrip": false,
32
+ "rstrip": false,
33
+ "normalized": false,
34
+ "special": true
35
+ },
36
+ {
37
+ "id": 185,
38
+ "content": "<pad>",
39
+ "single_word": false,
40
+ "lstrip": false,
41
+ "rstrip": false,
42
+ "normalized": false,
43
+ "special": true
44
+ }
45
+ ],
46
+ "normalizer": null,
47
+ "pre_tokenizer": {
48
+ "type": "Whitespace"
49
+ },
50
+ "post_processor": null,
51
+ "decoder": null,
52
+ "model": {
53
+ "type": "WordLevel",
54
+ "vocab": {
55
+ ",": 0,
56
+ "-1": 1,
57
+ "1linebegin": 2,
58
+ "1lineend": 3,
59
+ "1linemake": 4,
60
+ "<|endoftext|>": 5,
61
+ "=": 6,
62
+ "Coxeterreducedexpressionbegin": 7,
63
+ "Coxeterreducedexpressionend": 8,
64
+ "Coxeterreducedexpressionmake": 9,
65
+ "False": 10,
66
+ "Inversionvectorbegin": 11,
67
+ "Inversionvectorend": 12,
68
+ "Inversionvectormake": 13,
69
+ "Lehmercodebegin": 14,
70
+ "Lehmercodeend": 15,
71
+ "Lehmercodemake": 16,
72
+ "RSKtableauxbegin": 17,
73
+ "RSKtableauxend": 18,
74
+ "RSKtableauxmake": 19,
75
+ "True": 20,
76
+ "[": 21,
77
+ "]": 22,
78
+ "avoidsbegin": 23,
79
+ "avoidsend": 24,
80
+ "avoidsmake": 25,
81
+ "commutatormake": 26,
82
+ "complementmake": 27,
83
+ "conjugatemake": 28,
84
+ "cyclenotationbegin": 29,
85
+ "cyclenotationend": 30,
86
+ "cyclenotationmake": 31,
87
+ "cycletypebegin": 32,
88
+ "cycletypeend": 33,
89
+ "cycletypemake": 34,
90
+ "descentsbegin": 35,
91
+ "descentsend": 36,
92
+ "descentsmake": 37,
93
+ "descenttestbegin": 38,
94
+ "descenttestend": 39,
95
+ "exponentbegin": 40,
96
+ "exponentend": 41,
97
+ "fixedpointsbegin": 42,
98
+ "fixedpointsend": 43,
99
+ "fixedpointsmake": 44,
100
+ "in": 45,
101
+ "indexbegin": 46,
102
+ "indexend": 47,
103
+ "inversemake": 48,
104
+ "isderangementbegin": 49,
105
+ "isderangementend": 50,
106
+ "isderangementmake": 51,
107
+ "isevenbegin": 52,
108
+ "isevenend": 53,
109
+ "isevenmake": 54,
110
+ "isgrassmannianbegin": 55,
111
+ "isgrassmannianend": 56,
112
+ "isgrassmannianmake": 57,
113
+ "isinvolutionbegin": 58,
114
+ "isinvolutionend": 59,
115
+ "isinvolutionmake": 60,
116
+ "ldslengthbegin": 61,
117
+ "ldslengthend": 62,
118
+ "ldslengthmake": 63,
119
+ "leftBruhatbegin": 64,
120
+ "leftBruhatend": 65,
121
+ "leftBruhatmake": 66,
122
+ "leftdescenttestmake": 67,
123
+ "lengthbegin": 68,
124
+ "lengthend": 69,
125
+ "lengthmake": 70,
126
+ "lislengthbegin": 71,
127
+ "lislengthend": 72,
128
+ "lislengthmake": 73,
129
+ "longestincreasingsubsequencebegin": 74,
130
+ "longestincreasingsubsequenceend": 75,
131
+ "longestincreasingsubsequencemake": 76,
132
+ "majorindexbegin": 77,
133
+ "majorindexend": 78,
134
+ "majorindexmake": 79,
135
+ "multiplyleftsimake": 80,
136
+ "multiplyrightsimake": 81,
137
+ "n10": 82,
138
+ "n11": 83,
139
+ "n3": 84,
140
+ "n4": 85,
141
+ "n5": 86,
142
+ "n6": 87,
143
+ "n7": 88,
144
+ "n8": 89,
145
+ "n9": 90,
146
+ "nochain": 91,
147
+ "nopattern": 92,
148
+ "numcyclesbegin": 93,
149
+ "numcyclesend": 94,
150
+ "numcyclesmake": 95,
151
+ "numdescentsbegin": 96,
152
+ "numdescentsend": 97,
153
+ "numdescentsmake": 98,
154
+ "numexcedancesbegin": 99,
155
+ "numexcedancesend": 100,
156
+ "numexcedancesmake": 101,
157
+ "numfixedpointsbegin": 102,
158
+ "numfixedpointsend": 103,
159
+ "numfixedpointsmake": 104,
160
+ "numinversionsbegin": 105,
161
+ "numinversionsend": 106,
162
+ "numinversionsmake": 107,
163
+ "orderbegin": 108,
164
+ "orderend": 109,
165
+ "ordermake": 110,
166
+ "peaksbegin": 111,
167
+ "peaksend": 112,
168
+ "peaksmake": 113,
169
+ "powermake": 114,
170
+ "property": 115,
171
+ "recoilsbegin": 116,
172
+ "recoilsend": 117,
173
+ "recoilsmake": 118,
174
+ "relativeleftmake": 119,
175
+ "relativerightmake": 120,
176
+ "reversemake": 121,
177
+ "rightdescenttestmake": 122,
178
+ "rskshapebegin": 123,
179
+ "rskshapeend": 124,
180
+ "rskshapemake": 125,
181
+ "signbegin": 126,
182
+ "signend": 127,
183
+ "signmake": 128,
184
+ "timesmake": 129,
185
+ "witnessbegin": 130,
186
+ "witnessend": 131,
187
+ "0": 132,
188
+ "1": 133,
189
+ "10": 134,
190
+ "11": 135,
191
+ "12": 136,
192
+ "13": 137,
193
+ "14": 138,
194
+ "15": 139,
195
+ "16": 140,
196
+ "17": 141,
197
+ "18": 142,
198
+ "19": 143,
199
+ "2": 144,
200
+ "20": 145,
201
+ "21": 146,
202
+ "22": 147,
203
+ "23": 148,
204
+ "24": 149,
205
+ "25": 150,
206
+ "26": 151,
207
+ "27": 152,
208
+ "28": 153,
209
+ "29": 154,
210
+ "3": 155,
211
+ "30": 156,
212
+ "31": 157,
213
+ "32": 158,
214
+ "33": 159,
215
+ "34": 160,
216
+ "35": 161,
217
+ "36": 162,
218
+ "37": 163,
219
+ "38": 164,
220
+ "39": 165,
221
+ "4": 166,
222
+ "40": 167,
223
+ "41": 168,
224
+ "42": 169,
225
+ "43": 170,
226
+ "44": 171,
227
+ "45": 172,
228
+ "46": 173,
229
+ "47": 174,
230
+ "48": 175,
231
+ "49": 176,
232
+ "5": 177,
233
+ "50": 178,
234
+ "51": 179,
235
+ "6": 180,
236
+ "7": 181,
237
+ "8": 182,
238
+ "9": 183,
239
+ "<unk>": 184,
240
+ "<pad>": 185
241
+ },
242
+ "unk_token": "<unk>"
243
+ }
244
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "5": {
4
+ "content": "<|endoftext|>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "184": {
12
+ "content": "<unk>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "185": {
20
+ "content": "<pad>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ }
27
+ },
28
+ "bos_token": "<|endoftext|>",
29
+ "clean_up_tokenization_spaces": false,
30
+ "eos_token": "<|endoftext|>",
31
+ "extra_special_tokens": {},
32
+ "max_length": 1000,
33
+ "model_max_length": 1000,
34
+ "pad_to_multiple_of": null,
35
+ "pad_token": "<pad>",
36
+ "pad_token_type_id": 0,
37
+ "padding_side": "right",
38
+ "stride": 0,
39
+ "tokenizer_class": "PreTrainedTokenizerFast",
40
+ "truncation_side": "right",
41
+ "truncation_strategy": "longest_first",
42
+ "unk_token": "<unk>"
43
+ }