Hugging Face's logo

anudit
/
finetuned-gte-base

Sentence Similarity
sentence-transformers
ONNX
Safetensors
English
new
feature-extraction
Generated from Trainer
dataset_size:32833
loss:MatryoshkaLoss
loss:MultipleNegativesRankingLoss
custom_code
Eval Results
text-embeddings-inference
Inference Endpoints
Model card Files Community
anudit commited on
Commit
2f86976
·
verified ·
1 Parent(s): da0eab1

Upload 13 files

Browse files
Files changed (5) hide show
  1. README.md +247 -211
  2. config_sentence_transformers.json +2 -2
  3. model.onnx +1 -1
  4. model.safetensors +1 -1
  5. training_args.bin +1 -1
README.md CHANGED
@@ -1,6 +1,5 @@
1
  ---
2
  base_model: Alibaba-NLP/gte-base-en-v1.5
3
- datasets: []
4
  language:
5
  - en
6
  library_name: sentence-transformers
@@ -75,7 +74,7 @@ widget:
75
  - What lessons can be learned from the historical context of employee relations
76
  in large corporations?
77
  model-index:
78
- - name: Custom Embedding Test - Anudit Nagar
79
  results:
80
  - task:
81
  type: information-retrieval
@@ -85,49 +84,49 @@ model-index:
85
  type: dim_768
86
  metrics:
87
  - type: cosine_accuracy@1
88
- value: 0.7683027145599123
89
  name: Cosine Accuracy@1
90
  - type: cosine_accuracy@3
91
- value: 0.8755141211955032
92
  name: Cosine Accuracy@3
93
  - type: cosine_accuracy@5
94
- value: 0.9097888675623801
95
  name: Cosine Accuracy@5
96
  - type: cosine_accuracy@10
97
- value: 0.9465313956676721
98
  name: Cosine Accuracy@10
99
  - type: cosine_precision@1
100
- value: 0.7683027145599123
101
  name: Cosine Precision@1
102
  - type: cosine_precision@3
103
- value: 0.29183804039850103
104
  name: Cosine Precision@3
105
  - type: cosine_precision@5
106
- value: 0.18195777351247602
107
  name: Cosine Precision@5
108
  - type: cosine_precision@10
109
- value: 0.09465313956676721
110
  name: Cosine Precision@10
111
  - type: cosine_recall@1
112
- value: 0.7683027145599123
113
  name: Cosine Recall@1
114
  - type: cosine_recall@3
115
- value: 0.8755141211955032
116
  name: Cosine Recall@3
117
  - type: cosine_recall@5
118
- value: 0.9097888675623801
119
  name: Cosine Recall@5
120
  - type: cosine_recall@10
121
- value: 0.9465313956676721
122
  name: Cosine Recall@10
123
  - type: cosine_ndcg@10
124
- value: 0.8566925927271383
125
  name: Cosine Ndcg@10
126
  - type: cosine_mrr@10
127
- value: 0.8279207524340517
128
  name: Cosine Mrr@10
129
  - type: cosine_map@100
130
- value: 0.8302321946792381
131
  name: Cosine Map@100
132
  - task:
133
  type: information-retrieval
@@ -137,49 +136,49 @@ model-index:
137
  type: dim_512
138
  metrics:
139
  - type: cosine_accuracy@1
140
- value: 0.762818755141212
141
  name: Cosine Accuracy@1
142
  - type: cosine_accuracy@3
143
- value: 0.8700301617768028
144
  name: Cosine Accuracy@3
145
  - type: cosine_accuracy@5
146
- value: 0.9062242939402249
147
  name: Cosine Accuracy@5
148
  - type: cosine_accuracy@10
149
- value: 0.946257197696737
150
  name: Cosine Accuracy@10
151
  - type: cosine_precision@1
152
- value: 0.762818755141212
153
  name: Cosine Precision@1
154
  - type: cosine_precision@3
155
- value: 0.2900100539256009
156
  name: Cosine Precision@3
157
  - type: cosine_precision@5
158
- value: 0.18124485878804497
159
  name: Cosine Precision@5
160
  - type: cosine_precision@10
161
- value: 0.09462571976967371
162
  name: Cosine Precision@10
163
  - type: cosine_recall@1
164
- value: 0.762818755141212
165
  name: Cosine Recall@1
166
  - type: cosine_recall@3
167
- value: 0.8700301617768028
168
  name: Cosine Recall@3
169
  - type: cosine_recall@5
170
- value: 0.9062242939402249
171
  name: Cosine Recall@5
172
  - type: cosine_recall@10
173
- value: 0.946257197696737
174
  name: Cosine Recall@10
175
  - type: cosine_ndcg@10
176
- value: 0.8529743473843932
177
  name: Cosine Ndcg@10
178
  - type: cosine_mrr@10
179
- value: 0.8231949721667308
180
  name: Cosine Mrr@10
181
  - type: cosine_map@100
182
- value: 0.825407004380477
183
  name: Cosine Map@100
184
  - task:
185
  type: information-retrieval
@@ -189,49 +188,49 @@ model-index:
189
  type: dim_256
190
  metrics:
191
  - type: cosine_accuracy@1
192
- value: 0.762818755141212
193
  name: Cosine Accuracy@1
194
  - type: cosine_accuracy@3
195
- value: 0.8683849739511927
196
  name: Cosine Accuracy@3
197
  - type: cosine_accuracy@5
198
- value: 0.9015629284343296
199
  name: Cosine Accuracy@5
200
  - type: cosine_accuracy@10
201
- value: 0.9418700301617768
202
  name: Cosine Accuracy@10
203
  - type: cosine_precision@1
204
- value: 0.762818755141212
205
  name: Cosine Precision@1
206
  - type: cosine_precision@3
207
- value: 0.28946165798373086
208
  name: Cosine Precision@3
209
  - type: cosine_precision@5
210
- value: 0.18031258568686592
211
  name: Cosine Precision@5
212
  - type: cosine_precision@10
213
- value: 0.09418700301617768
214
  name: Cosine Precision@10
215
  - type: cosine_recall@1
216
- value: 0.762818755141212
217
  name: Cosine Recall@1
218
  - type: cosine_recall@3
219
- value: 0.8683849739511927
220
  name: Cosine Recall@3
221
  - type: cosine_recall@5
222
- value: 0.9015629284343296
223
  name: Cosine Recall@5
224
  - type: cosine_recall@10
225
- value: 0.9418700301617768
226
  name: Cosine Recall@10
227
  - type: cosine_ndcg@10
228
- value: 0.850685453111757
229
  name: Cosine Ndcg@10
230
  - type: cosine_mrr@10
231
- value: 0.8215859088357048
232
  name: Cosine Mrr@10
233
  - type: cosine_map@100
234
- value: 0.8239714751253995
235
  name: Cosine Map@100
236
  - task:
237
  type: information-retrieval
@@ -241,49 +240,49 @@ model-index:
241
  type: dim_128
242
  metrics:
243
  - type: cosine_accuracy@1
244
- value: 0.7573347957225116
245
  name: Cosine Accuracy@1
246
  - type: cosine_accuracy@3
247
- value: 0.8634494104743625
248
  name: Cosine Accuracy@3
249
  - type: cosine_accuracy@5
250
- value: 0.8952563751028242
251
  name: Cosine Accuracy@5
252
  - type: cosine_accuracy@10
253
- value: 0.9347408829174664
254
  name: Cosine Accuracy@10
255
  - type: cosine_precision@1
256
- value: 0.7573347957225116
257
  name: Cosine Precision@1
258
  - type: cosine_precision@3
259
- value: 0.2878164701581208
260
  name: Cosine Precision@3
261
  - type: cosine_precision@5
262
- value: 0.17905127502056484
263
  name: Cosine Precision@5
264
  - type: cosine_precision@10
265
- value: 0.09347408829174664
266
  name: Cosine Precision@10
267
  - type: cosine_recall@1
268
- value: 0.7573347957225116
269
  name: Cosine Recall@1
270
  - type: cosine_recall@3
271
- value: 0.8634494104743625
272
  name: Cosine Recall@3
273
  - type: cosine_recall@5
274
- value: 0.8952563751028242
275
  name: Cosine Recall@5
276
  - type: cosine_recall@10
277
- value: 0.9347408829174664
278
  name: Cosine Recall@10
279
  - type: cosine_ndcg@10
280
- value: 0.8445055968214926
281
  name: Cosine Ndcg@10
282
  - type: cosine_mrr@10
283
- value: 0.8157123053956075
284
  name: Cosine Mrr@10
285
  - type: cosine_map@100
286
- value: 0.8184088689781863
287
  name: Cosine Map@100
288
  - task:
289
  type: information-retrieval
@@ -293,55 +292,55 @@ model-index:
293
  type: dim_64
294
  metrics:
295
  - type: cosine_accuracy@1
296
- value: 0.7419797093501508
297
  name: Cosine Accuracy@1
298
  - type: cosine_accuracy@3
299
- value: 0.8530298875788319
300
  name: Cosine Accuracy@3
301
  - type: cosine_accuracy@5
302
- value: 0.8859336440910337
303
  name: Cosine Accuracy@5
304
  - type: cosine_accuracy@10
305
- value: 0.9284343295859611
306
  name: Cosine Accuracy@10
307
  - type: cosine_precision@1
308
- value: 0.7419797093501508
309
  name: Cosine Precision@1
310
  - type: cosine_precision@3
311
- value: 0.28434329585961066
312
  name: Cosine Precision@3
313
  - type: cosine_precision@5
314
- value: 0.17718672881820677
315
  name: Cosine Precision@5
316
  - type: cosine_precision@10
317
- value: 0.09284343295859611
318
  name: Cosine Precision@10
319
  - type: cosine_recall@1
320
- value: 0.7419797093501508
321
  name: Cosine Recall@1
322
  - type: cosine_recall@3
323
- value: 0.8530298875788319
324
  name: Cosine Recall@3
325
  - type: cosine_recall@5
326
- value: 0.8859336440910337
327
  name: Cosine Recall@5
328
  - type: cosine_recall@10
329
- value: 0.9284343295859611
330
  name: Cosine Recall@10
331
  - type: cosine_ndcg@10
332
- value: 0.8334906130922063
333
  name: Cosine Ndcg@10
334
  - type: cosine_mrr@10
335
- value: 0.8032139919307455
336
  name: Cosine Mrr@10
337
  - type: cosine_map@100
338
- value: 0.8057146368194794
339
  name: Cosine Map@100
340
  ---
341
 
342
- # Custom Embedding Test - Anudit Nagar
343
 
344
- This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Alibaba-NLP/gte-base-en-v1.5](https://huggingface.co/Alibaba-NLP/gte-base-en-v1.5). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
345
 
346
  ## Model Details
347
 
@@ -351,7 +350,8 @@ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [A
351
  - **Maximum Sequence Length:** 8192 tokens
352
  - **Output Dimensionality:** 768 tokens
353
  - **Similarity Function:** Cosine Similarity
354
- <!-- - **Training Dataset:** Unknown -->
 
355
  - **Language:** en
356
  - **License:** apache-2.0
357
 
@@ -436,21 +436,21 @@ You can finetune this model on your own dataset.
436
 
437
  | Metric | Value |
438
  |:--------------------|:-----------|
439
- | cosine_accuracy@1 | 0.7683 |
440
- | cosine_accuracy@3 | 0.8755 |
441
- | cosine_accuracy@5 | 0.9098 |
442
- | cosine_accuracy@10 | 0.9465 |
443
- | cosine_precision@1 | 0.7683 |
444
- | cosine_precision@3 | 0.2918 |
445
- | cosine_precision@5 | 0.182 |
446
- | cosine_precision@10 | 0.0947 |
447
- | cosine_recall@1 | 0.7683 |
448
- | cosine_recall@3 | 0.8755 |
449
- | cosine_recall@5 | 0.9098 |
450
- | cosine_recall@10 | 0.9465 |
451
- | cosine_ndcg@10 | 0.8567 |
452
- | cosine_mrr@10 | 0.8279 |
453
- | **cosine_map@100** | **0.8302** |
454
 
455
  #### Information Retrieval
456
  * Dataset: `dim_512`
@@ -458,43 +458,43 @@ You can finetune this model on your own dataset.
458
 
459
  | Metric | Value |
460
  |:--------------------|:-----------|
461
- | cosine_accuracy@1 | 0.7628 |
462
- | cosine_accuracy@3 | 0.87 |
463
- | cosine_accuracy@5 | 0.9062 |
464
- | cosine_accuracy@10 | 0.9463 |
465
- | cosine_precision@1 | 0.7628 |
466
- | cosine_precision@3 | 0.29 |
467
- | cosine_precision@5 | 0.1812 |
468
- | cosine_precision@10 | 0.0946 |
469
- | cosine_recall@1 | 0.7628 |
470
- | cosine_recall@3 | 0.87 |
471
- | cosine_recall@5 | 0.9062 |
472
- | cosine_recall@10 | 0.9463 |
473
- | cosine_ndcg@10 | 0.853 |
474
- | cosine_mrr@10 | 0.8232 |
475
- | **cosine_map@100** | **0.8254** |
476
 
477
  #### Information Retrieval
478
  * Dataset: `dim_256`
479
  * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
480
 
481
- | Metric | Value |
482
- |:--------------------|:----------|
483
- | cosine_accuracy@1 | 0.7628 |
484
- | cosine_accuracy@3 | 0.8684 |
485
- | cosine_accuracy@5 | 0.9016 |
486
- | cosine_accuracy@10 | 0.9419 |
487
- | cosine_precision@1 | 0.7628 |
488
- | cosine_precision@3 | 0.2895 |
489
- | cosine_precision@5 | 0.1803 |
490
- | cosine_precision@10 | 0.0942 |
491
- | cosine_recall@1 | 0.7628 |
492
- | cosine_recall@3 | 0.8684 |
493
- | cosine_recall@5 | 0.9016 |
494
- | cosine_recall@10 | 0.9419 |
495
- | cosine_ndcg@10 | 0.8507 |
496
- | cosine_mrr@10 | 0.8216 |
497
- | **cosine_map@100** | **0.824** |
498
 
499
  #### Information Retrieval
500
  * Dataset: `dim_128`
@@ -502,21 +502,21 @@ You can finetune this model on your own dataset.
502
 
503
  | Metric | Value |
504
  |:--------------------|:-----------|
505
- | cosine_accuracy@1 | 0.7573 |
506
- | cosine_accuracy@3 | 0.8634 |
507
- | cosine_accuracy@5 | 0.8953 |
508
- | cosine_accuracy@10 | 0.9347 |
509
- | cosine_precision@1 | 0.7573 |
510
- | cosine_precision@3 | 0.2878 |
511
- | cosine_precision@5 | 0.1791 |
512
- | cosine_precision@10 | 0.0935 |
513
- | cosine_recall@1 | 0.7573 |
514
- | cosine_recall@3 | 0.8634 |
515
- | cosine_recall@5 | 0.8953 |
516
- | cosine_recall@10 | 0.9347 |
517
- | cosine_ndcg@10 | 0.8445 |
518
- | cosine_mrr@10 | 0.8157 |
519
- | **cosine_map@100** | **0.8184** |
520
 
521
  #### Information Retrieval
522
  * Dataset: `dim_64`
@@ -524,21 +524,21 @@ You can finetune this model on your own dataset.
524
 
525
  | Metric | Value |
526
  |:--------------------|:-----------|
527
- | cosine_accuracy@1 | 0.742 |
528
- | cosine_accuracy@3 | 0.853 |
529
- | cosine_accuracy@5 | 0.8859 |
530
- | cosine_accuracy@10 | 0.9284 |
531
- | cosine_precision@1 | 0.742 |
532
- | cosine_precision@3 | 0.2843 |
533
- | cosine_precision@5 | 0.1772 |
534
- | cosine_precision@10 | 0.0928 |
535
- | cosine_recall@1 | 0.742 |
536
- | cosine_recall@3 | 0.853 |
537
- | cosine_recall@5 | 0.8859 |
538
- | cosine_recall@10 | 0.9284 |
539
- | cosine_ndcg@10 | 0.8335 |
540
- | cosine_mrr@10 | 0.8032 |
541
- | **cosine_map@100** | **0.8057** |
542
 
543
  <!--
544
  ## Bias, Risks and Limitations
@@ -556,9 +556,9 @@ You can finetune this model on your own dataset.
556
 
557
  ### Training Dataset
558
 
559
- #### Unnamed Dataset
560
-
561
 
 
562
  * Size: 32,833 training samples
563
  * Columns: <code>positive</code> and <code>anchor</code>
564
  * Approximate statistics based on the first 1000 samples:
@@ -598,11 +598,11 @@ You can finetune this model on your own dataset.
598
  #### Non-Default Hyperparameters
599
 
600
  - `eval_strategy`: epoch
601
- - `per_device_train_batch_size`: 32
602
- - `per_device_eval_batch_size`: 16
603
- - `gradient_accumulation_steps`: 16
604
- - `learning_rate`: 0.0002
605
- - `num_train_epochs`: 5
606
  - `lr_scheduler_type`: cosine
607
  - `warmup_ratio`: 0.1
608
  - `bf16`: True
@@ -616,20 +616,20 @@ You can finetune this model on your own dataset.
616
  - `do_predict`: False
617
  - `eval_strategy`: epoch
618
  - `prediction_loss_only`: True
619
- - `per_device_train_batch_size`: 32
620
- - `per_device_eval_batch_size`: 16
621
  - `per_gpu_train_batch_size`: None
622
  - `per_gpu_eval_batch_size`: None
623
- - `gradient_accumulation_steps`: 16
624
  - `eval_accumulation_steps`: None
625
  - `torch_empty_cache_steps`: None
626
- - `learning_rate`: 0.0002
627
  - `weight_decay`: 0.0
628
  - `adam_beta1`: 0.9
629
  - `adam_beta2`: 0.999
630
  - `adam_epsilon`: 1e-08
631
  - `max_grad_norm`: 1.0
632
- - `num_train_epochs`: 5
633
  - `max_steps`: -1
634
  - `lr_scheduler_type`: cosine
635
  - `lr_scheduler_kwargs`: {}
@@ -727,52 +727,88 @@ You can finetune this model on your own dataset.
727
  </details>
728
 
729
  ### Training Logs
730
- | Epoch | Step | Training Loss | dim_128_cosine_map@100 | dim_256_cosine_map@100 | dim_512_cosine_map@100 | dim_64_cosine_map@100 | dim_768_cosine_map@100 |
731
- |:----------:|:------:|:-------------:|:----------------------:|:----------------------:|:----------------------:|:---------------------:|:----------------------:|
732
- | 0.1558 | 10 | 0.7195 | - | - | - | - | - |
733
- | 0.3116 | 20 | 0.324 | - | - | - | - | - |
734
- | 0.4674 | 30 | 0.238 | - | - | - | - | - |
735
- | 0.6232 | 40 | 0.2265 | - | - | - | - | - |
736
- | 0.7790 | 50 | 0.1825 | - | - | - | - | - |
737
- | 0.9348 | 60 | 0.1938 | - | - | - | - | - |
738
- | **0.9971** | **64** | **-** | **0.8054** | **0.8198** | **0.8276** | **0.7796** | **0.8329** |
739
- | 1.0906 | 70 | 0.1397 | - | - | - | - | - |
740
- | 1.2463 | 80 | 0.0611 | - | - | - | - | - |
741
- | 1.4021 | 90 | 0.0506 | - | - | - | - | - |
742
- | 1.5579 | 100 | 0.047 | - | - | - | - | - |
743
- | 1.7137 | 110 | 0.0327 | - | - | - | - | - |
744
- | 1.8695 | 120 | 0.034 | - | - | - | - | - |
745
- | 1.9942 | 128 | - | 0.8036 | 0.8135 | 0.8187 | 0.7861 | 0.8243 |
746
- | 2.0253 | 130 | 0.0319 | - | - | - | - | - |
747
- | 2.1811 | 140 | 0.0347 | - | - | - | - | - |
748
- | 2.3369 | 150 | 0.021 | - | - | - | - | - |
749
- | 2.4927 | 160 | 0.0169 | - | - | - | - | - |
750
- | 2.6485 | 170 | 0.0135 | - | - | - | - | - |
751
- | 2.8043 | 180 | 0.0123 | - | - | - | - | - |
752
- | 2.9601 | 190 | 0.0111 | - | - | - | - | - |
753
- | 2.9912 | 192 | - | 0.8109 | 0.8179 | 0.8213 | 0.7973 | 0.8264 |
754
- | 3.1159 | 200 | 0.0083 | - | - | - | - | - |
755
- | 3.2717 | 210 | 0.0088 | - | - | - | - | - |
756
- | 3.4275 | 220 | 0.005 | - | - | - | - | - |
757
- | 3.5833 | 230 | 0.005 | - | - | - | - | - |
758
- | 3.7390 | 240 | 0.0043 | - | - | - | - | - |
759
- | 3.8948 | 250 | 0.0058 | - | - | - | - | - |
760
- | 3.9883 | 256 | - | 0.8163 | 0.8244 | 0.8260 | 0.8045 | 0.8287 |
761
- | 4.0506 | 260 | 0.0057 | - | - | - | - | - |
762
- | 4.2064 | 270 | 0.0035 | - | - | - | - | - |
763
- | 4.3622 | 280 | 0.0033 | - | - | - | - | - |
764
- | 4.5180 | 290 | 0.0032 | - | - | - | - | - |
765
- | 4.6738 | 300 | 0.0031 | - | - | - | - | - |
766
- | 4.8296 | 310 | 0.0038 | - | - | - | - | - |
767
- | 4.9854 | 320 | 0.0042 | 0.8184 | 0.8240 | 0.8254 | 0.8057 | 0.8302 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
768
 
769
  * The bold row denotes the saved checkpoint.
770
 
771
  ### Framework Versions
772
  - Python: 3.12.5
773
- - Sentence Transformers: 3.0.1
774
  - Transformers: 4.44.2
775
- - PyTorch: 2.4.0
776
  - Accelerate: 0.33.0
777
  - Datasets: 2.21.0
778
  - Tokenizers: 0.19.1
@@ -797,7 +833,7 @@ You can finetune this model on your own dataset.
797
  #### MatryoshkaLoss
798
  ```bibtex
799
  @misc{kusupati2024matryoshka,
800
- title={Matryoshka Representation Learning},
801
  author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
802
  year={2024},
803
  eprint={2205.13147},
@@ -809,7 +845,7 @@ You can finetune this model on your own dataset.
809
  #### MultipleNegativesRankingLoss
810
  ```bibtex
811
  @misc{henderson2017efficient,
812
- title={Efficient Natural Language Response Suggestion for Smart Reply},
813
  author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
814
  year={2017},
815
  eprint={1705.00652},
 
1
  ---
2
  base_model: Alibaba-NLP/gte-base-en-v1.5
 
3
  language:
4
  - en
5
  library_name: sentence-transformers
 
74
  - What lessons can be learned from the historical context of employee relations
75
  in large corporations?
76
  model-index:
77
+ - name: Alchemy Embedding - Anudit Nagar
78
  results:
79
  - task:
80
  type: information-retrieval
 
84
  type: dim_768
85
  metrics:
86
  - type: cosine_accuracy@1
87
+ value: 0.782012613106663
88
  name: Cosine Accuracy@1
89
  - type: cosine_accuracy@3
90
+ value: 0.8889498217713189
91
  name: Cosine Accuracy@3
92
  - type: cosine_accuracy@5
93
+ value: 0.9248697559638058
94
  name: Cosine Accuracy@5
95
  - type: cosine_accuracy@10
96
+ value: 0.9520153550863724
97
  name: Cosine Accuracy@10
98
  - type: cosine_precision@1
99
+ value: 0.782012613106663
100
  name: Cosine Precision@1
101
  - type: cosine_precision@3
102
+ value: 0.29631660725710623
103
  name: Cosine Precision@3
104
  - type: cosine_precision@5
105
+ value: 0.1849739511927612
106
  name: Cosine Precision@5
107
  - type: cosine_precision@10
108
+ value: 0.09520153550863725
109
  name: Cosine Precision@10
110
  - type: cosine_recall@1
111
+ value: 0.782012613106663
112
  name: Cosine Recall@1
113
  - type: cosine_recall@3
114
+ value: 0.8889498217713189
115
  name: Cosine Recall@3
116
  - type: cosine_recall@5
117
+ value: 0.9248697559638058
118
  name: Cosine Recall@5
119
  - type: cosine_recall@10
120
+ value: 0.9520153550863724
121
  name: Cosine Recall@10
122
  - type: cosine_ndcg@10
123
+ value: 0.867555587052628
124
  name: Cosine Ndcg@10
125
  - type: cosine_mrr@10
126
+ value: 0.8402608580220322
127
  name: Cosine Mrr@10
128
  - type: cosine_map@100
129
+ value: 0.8422322227138224
130
  name: Cosine Map@100
131
  - task:
132
  type: information-retrieval
 
136
  type: dim_512
137
  metrics:
138
  - type: cosine_accuracy@1
139
+ value: 0.780367425281053
140
  name: Cosine Accuracy@1
141
  - type: cosine_accuracy@3
142
+ value: 0.8848368522072937
143
  name: Cosine Accuracy@3
144
  - type: cosine_accuracy@5
145
+ value: 0.9221277762544557
146
  name: Cosine Accuracy@5
147
  - type: cosine_accuracy@10
148
+ value: 0.9514669591445023
149
  name: Cosine Accuracy@10
150
  - type: cosine_precision@1
151
+ value: 0.780367425281053
152
  name: Cosine Precision@1
153
  - type: cosine_precision@3
154
+ value: 0.2949456174024312
155
  name: Cosine Precision@3
156
  - type: cosine_precision@5
157
+ value: 0.1844255552508912
158
  name: Cosine Precision@5
159
  - type: cosine_precision@10
160
+ value: 0.09514669591445023
161
  name: Cosine Precision@10
162
  - type: cosine_recall@1
163
+ value: 0.780367425281053
164
  name: Cosine Recall@1
165
  - type: cosine_recall@3
166
+ value: 0.8848368522072937
167
  name: Cosine Recall@3
168
  - type: cosine_recall@5
169
+ value: 0.9221277762544557
170
  name: Cosine Recall@5
171
  - type: cosine_recall@10
172
+ value: 0.9514669591445023
173
  name: Cosine Recall@10
174
  - type: cosine_ndcg@10
175
+ value: 0.8661558392165704
176
  name: Cosine Ndcg@10
177
  - type: cosine_mrr@10
178
+ value: 0.838656038231032
179
  name: Cosine Mrr@10
180
  - type: cosine_map@100
181
+ value: 0.8405372438205077
182
  name: Cosine Map@100
183
  - task:
184
  type: information-retrieval
 
188
  type: dim_256
189
  metrics:
190
  - type: cosine_accuracy@1
191
+ value: 0.7754318618042226
192
  name: Cosine Accuracy@1
193
  - type: cosine_accuracy@3
194
+ value: 0.8804496846723334
195
  name: Cosine Accuracy@3
196
  - type: cosine_accuracy@5
197
+ value: 0.9169180148066904
198
  name: Cosine Accuracy@5
199
  - type: cosine_accuracy@10
200
+ value: 0.9468055936386071
201
  name: Cosine Accuracy@10
202
  - type: cosine_precision@1
203
+ value: 0.7754318618042226
204
  name: Cosine Precision@1
205
  - type: cosine_precision@3
206
+ value: 0.2934832282241111
207
  name: Cosine Precision@3
208
  - type: cosine_precision@5
209
+ value: 0.18338360296133807
210
  name: Cosine Precision@5
211
  - type: cosine_precision@10
212
+ value: 0.09468055936386072
213
  name: Cosine Precision@10
214
  - type: cosine_recall@1
215
+ value: 0.7754318618042226
216
  name: Cosine Recall@1
217
  - type: cosine_recall@3
218
+ value: 0.8804496846723334
219
  name: Cosine Recall@3
220
  - type: cosine_recall@5
221
+ value: 0.9169180148066904
222
  name: Cosine Recall@5
223
  - type: cosine_recall@10
224
+ value: 0.9468055936386071
225
  name: Cosine Recall@10
226
  - type: cosine_ndcg@10
227
+ value: 0.8613819477350178
228
  name: Cosine Ndcg@10
229
  - type: cosine_mrr@10
230
+ value: 0.8338379881703168
231
  name: Cosine Mrr@10
232
  - type: cosine_map@100
233
+ value: 0.8360735900013385
234
  name: Cosine Map@100
235
  - task:
236
  type: information-retrieval
 
240
  type: dim_128
241
  metrics:
242
  - type: cosine_accuracy@1
243
+ value: 0.7617219632574719
244
  name: Cosine Accuracy@1
245
  - type: cosine_accuracy@3
246
+ value: 0.871675349602413
247
  name: Cosine Accuracy@3
248
  - type: cosine_accuracy@5
249
+ value: 0.9117082533589251
250
  name: Cosine Accuracy@5
251
  - type: cosine_accuracy@10
252
+ value: 0.9418700301617768
253
  name: Cosine Accuracy@10
254
  - type: cosine_precision@1
255
+ value: 0.7617219632574719
256
  name: Cosine Precision@1
257
  - type: cosine_precision@3
258
+ value: 0.2905584498674709
259
  name: Cosine Precision@3
260
  - type: cosine_precision@5
261
+ value: 0.18234165067178504
262
  name: Cosine Precision@5
263
  - type: cosine_precision@10
264
+ value: 0.09418700301617768
265
  name: Cosine Precision@10
266
  - type: cosine_recall@1
267
+ value: 0.7617219632574719
268
  name: Cosine Recall@1
269
  - type: cosine_recall@3
270
+ value: 0.871675349602413
271
  name: Cosine Recall@3
272
  - type: cosine_recall@5
273
+ value: 0.9117082533589251
274
  name: Cosine Recall@5
275
  - type: cosine_recall@10
276
+ value: 0.9418700301617768
277
  name: Cosine Recall@10
278
  - type: cosine_ndcg@10
279
+ value: 0.851649908463093
280
  name: Cosine Ndcg@10
281
  - type: cosine_mrr@10
282
+ value: 0.8225671458602635
283
  name: Cosine Mrr@10
284
  - type: cosine_map@100
285
+ value: 0.8248455884524328
286
  name: Cosine Map@100
287
  - task:
288
  type: information-retrieval
 
292
  type: dim_64
293
  metrics:
294
  - type: cosine_accuracy@1
295
+ value: 0.7408829174664108
296
  name: Cosine Accuracy@1
297
  - type: cosine_accuracy@3
298
+ value: 0.853852481491637
299
  name: Cosine Accuracy@3
300
  - type: cosine_accuracy@5
301
+ value: 0.8936111872772141
302
  name: Cosine Accuracy@5
303
  - type: cosine_accuracy@10
304
+ value: 0.9292569234987661
305
  name: Cosine Accuracy@10
306
  - type: cosine_precision@1
307
+ value: 0.7408829174664108
308
  name: Cosine Precision@1
309
  - type: cosine_precision@3
310
+ value: 0.28461749383054563
311
  name: Cosine Precision@3
312
  - type: cosine_precision@5
313
+ value: 0.17872223745544283
314
  name: Cosine Precision@5
315
  - type: cosine_precision@10
316
+ value: 0.0929256923498766
317
  name: Cosine Precision@10
318
  - type: cosine_recall@1
319
+ value: 0.7408829174664108
320
  name: Cosine Recall@1
321
  - type: cosine_recall@3
322
+ value: 0.853852481491637
323
  name: Cosine Recall@3
324
  - type: cosine_recall@5
325
+ value: 0.8936111872772141
326
  name: Cosine Recall@5
327
  - type: cosine_recall@10
328
+ value: 0.9292569234987661
329
  name: Cosine Recall@10
330
  - type: cosine_ndcg@10
331
+ value: 0.8338956659320366
332
  name: Cosine Ndcg@10
333
  - type: cosine_mrr@10
334
+ value: 0.8033378162525404
335
  name: Cosine Mrr@10
336
  - type: cosine_map@100
337
+ value: 0.8057702637208689
338
  name: Cosine Map@100
339
  ---
340
 
341
+ # Alchemy Embedding - Anudit Nagar
342
 
343
+ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Alibaba-NLP/gte-base-en-v1.5](https://huggingface.co/Alibaba-NLP/gte-base-en-v1.5) on the json dataset. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
344
 
345
  ## Model Details
346
 
 
350
  - **Maximum Sequence Length:** 8192 tokens
351
  - **Output Dimensionality:** 768 tokens
352
  - **Similarity Function:** Cosine Similarity
353
+ - **Training Dataset:**
354
+ - json
355
  - **Language:** en
356
  - **License:** apache-2.0
357
 
 
436
 
437
  | Metric | Value |
438
  |:--------------------|:-----------|
439
+ | cosine_accuracy@1 | 0.782 |
440
+ | cosine_accuracy@3 | 0.8889 |
441
+ | cosine_accuracy@5 | 0.9249 |
442
+ | cosine_accuracy@10 | 0.952 |
443
+ | cosine_precision@1 | 0.782 |
444
+ | cosine_precision@3 | 0.2963 |
445
+ | cosine_precision@5 | 0.185 |
446
+ | cosine_precision@10 | 0.0952 |
447
+ | cosine_recall@1 | 0.782 |
448
+ | cosine_recall@3 | 0.8889 |
449
+ | cosine_recall@5 | 0.9249 |
450
+ | cosine_recall@10 | 0.952 |
451
+ | cosine_ndcg@10 | 0.8676 |
452
+ | cosine_mrr@10 | 0.8403 |
453
+ | **cosine_map@100** | **0.8422** |
454
 
455
  #### Information Retrieval
456
  * Dataset: `dim_512`
 
458
 
459
  | Metric | Value |
460
  |:--------------------|:-----------|
461
+ | cosine_accuracy@1 | 0.7804 |
462
+ | cosine_accuracy@3 | 0.8848 |
463
+ | cosine_accuracy@5 | 0.9221 |
464
+ | cosine_accuracy@10 | 0.9515 |
465
+ | cosine_precision@1 | 0.7804 |
466
+ | cosine_precision@3 | 0.2949 |
467
+ | cosine_precision@5 | 0.1844 |
468
+ | cosine_precision@10 | 0.0951 |
469
+ | cosine_recall@1 | 0.7804 |
470
+ | cosine_recall@3 | 0.8848 |
471
+ | cosine_recall@5 | 0.9221 |
472
+ | cosine_recall@10 | 0.9515 |
473
+ | cosine_ndcg@10 | 0.8662 |
474
+ | cosine_mrr@10 | 0.8387 |
475
+ | **cosine_map@100** | **0.8405** |
476
 
477
  #### Information Retrieval
478
  * Dataset: `dim_256`
479
  * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
480
 
481
+ | Metric | Value |
482
+ |:--------------------|:-----------|
483
+ | cosine_accuracy@1 | 0.7754 |
484
+ | cosine_accuracy@3 | 0.8804 |
485
+ | cosine_accuracy@5 | 0.9169 |
486
+ | cosine_accuracy@10 | 0.9468 |
487
+ | cosine_precision@1 | 0.7754 |
488
+ | cosine_precision@3 | 0.2935 |
489
+ | cosine_precision@5 | 0.1834 |
490
+ | cosine_precision@10 | 0.0947 |
491
+ | cosine_recall@1 | 0.7754 |
492
+ | cosine_recall@3 | 0.8804 |
493
+ | cosine_recall@5 | 0.9169 |
494
+ | cosine_recall@10 | 0.9468 |
495
+ | cosine_ndcg@10 | 0.8614 |
496
+ | cosine_mrr@10 | 0.8338 |
497
+ | **cosine_map@100** | **0.8361** |
498
 
499
  #### Information Retrieval
500
  * Dataset: `dim_128`
 
502
 
503
  | Metric | Value |
504
  |:--------------------|:-----------|
505
+ | cosine_accuracy@1 | 0.7617 |
506
+ | cosine_accuracy@3 | 0.8717 |
507
+ | cosine_accuracy@5 | 0.9117 |
508
+ | cosine_accuracy@10 | 0.9419 |
509
+ | cosine_precision@1 | 0.7617 |
510
+ | cosine_precision@3 | 0.2906 |
511
+ | cosine_precision@5 | 0.1823 |
512
+ | cosine_precision@10 | 0.0942 |
513
+ | cosine_recall@1 | 0.7617 |
514
+ | cosine_recall@3 | 0.8717 |
515
+ | cosine_recall@5 | 0.9117 |
516
+ | cosine_recall@10 | 0.9419 |
517
+ | cosine_ndcg@10 | 0.8516 |
518
+ | cosine_mrr@10 | 0.8226 |
519
+ | **cosine_map@100** | **0.8248** |
520
 
521
  #### Information Retrieval
522
  * Dataset: `dim_64`
 
524
 
525
  | Metric | Value |
526
  |:--------------------|:-----------|
527
+ | cosine_accuracy@1 | 0.7409 |
528
+ | cosine_accuracy@3 | 0.8539 |
529
+ | cosine_accuracy@5 | 0.8936 |
530
+ | cosine_accuracy@10 | 0.9293 |
531
+ | cosine_precision@1 | 0.7409 |
532
+ | cosine_precision@3 | 0.2846 |
533
+ | cosine_precision@5 | 0.1787 |
534
+ | cosine_precision@10 | 0.0929 |
535
+ | cosine_recall@1 | 0.7409 |
536
+ | cosine_recall@3 | 0.8539 |
537
+ | cosine_recall@5 | 0.8936 |
538
+ | cosine_recall@10 | 0.9293 |
539
+ | cosine_ndcg@10 | 0.8339 |
540
+ | cosine_mrr@10 | 0.8033 |
541
+ | **cosine_map@100** | **0.8058** |
542
 
543
  <!--
544
  ## Bias, Risks and Limitations
 
556
 
557
  ### Training Dataset
558
 
559
+ #### json
 
560
 
561
+ * Dataset: json
562
  * Size: 32,833 training samples
563
  * Columns: <code>positive</code> and <code>anchor</code>
564
  * Approximate statistics based on the first 1000 samples:
 
598
  #### Non-Default Hyperparameters
599
 
600
  - `eval_strategy`: epoch
601
+ - `per_device_train_batch_size`: 24
602
+ - `per_device_eval_batch_size`: 24
603
+ - `gradient_accumulation_steps`: 8
604
+ - `learning_rate`: 2e-05
605
+ - `num_train_epochs`: 4
606
  - `lr_scheduler_type`: cosine
607
  - `warmup_ratio`: 0.1
608
  - `bf16`: True
 
616
  - `do_predict`: False
617
  - `eval_strategy`: epoch
618
  - `prediction_loss_only`: True
619
+ - `per_device_train_batch_size`: 24
620
+ - `per_device_eval_batch_size`: 24
621
  - `per_gpu_train_batch_size`: None
622
  - `per_gpu_eval_batch_size`: None
623
+ - `gradient_accumulation_steps`: 8
624
  - `eval_accumulation_steps`: None
625
  - `torch_empty_cache_steps`: None
626
+ - `learning_rate`: 2e-05
627
  - `weight_decay`: 0.0
628
  - `adam_beta1`: 0.9
629
  - `adam_beta2`: 0.999
630
  - `adam_epsilon`: 1e-08
631
  - `max_grad_norm`: 1.0
632
+ - `num_train_epochs`: 4
633
  - `max_steps`: -1
634
  - `lr_scheduler_type`: cosine
635
  - `lr_scheduler_kwargs`: {}
 
727
  </details>
728
 
729
  ### Training Logs
730
+ | Epoch | Step | Training Loss | dim_128_cosine_map@100 | dim_256_cosine_map@100 | dim_512_cosine_map@100 | dim_64_cosine_map@100 | dim_768_cosine_map@100 |
731
+ |:----------:|:-------:|:-------------:|:----------------------:|:----------------------:|:----------------------:|:---------------------:|:----------------------:|
732
+ | 0.0584 | 10 | 0.8567 | - | - | - | - | - |
733
+ | 0.1169 | 20 | 0.6549 | - | - | - | - | - |
734
+ | 0.1753 | 30 | 0.5407 | - | - | - | - | - |
735
+ | 0.2337 | 40 | 0.4586 | - | - | - | - | - |
736
+ | 0.2922 | 50 | 0.3914 | - | - | - | - | - |
737
+ | 0.3506 | 60 | 0.4104 | - | - | - | - | - |
738
+ | 0.4091 | 70 | 0.299 | - | - | - | - | - |
739
+ | 0.4675 | 80 | 0.2444 | - | - | - | - | - |
740
+ | 0.5259 | 90 | 0.2367 | - | - | - | - | - |
741
+ | 0.5844 | 100 | 0.2302 | - | - | - | - | - |
742
+ | 0.6428 | 110 | 0.2356 | - | - | - | - | - |
743
+ | 0.7012 | 120 | 0.1537 | - | - | - | - | - |
744
+ | 0.7597 | 130 | 0.2043 | - | - | - | - | - |
745
+ | 0.8181 | 140 | 0.1606 | - | - | - | - | - |
746
+ | 0.8766 | 150 | 0.1896 | - | - | - | - | - |
747
+ | 0.9350 | 160 | 0.1766 | - | - | - | - | - |
748
+ | 0.9934 | 170 | 0.1259 | - | - | - | - | - |
749
+ | 0.9993 | 171 | - | 0.8115 | 0.8233 | 0.8321 | 0.7829 | 0.8340 |
750
+ | 1.0519 | 180 | 0.1661 | - | - | - | - | - |
751
+ | 1.1103 | 190 | 0.1632 | - | - | - | - | - |
752
+ | 1.1687 | 200 | 0.1032 | - | - | - | - | - |
753
+ | 1.2272 | 210 | 0.1037 | - | - | - | - | - |
754
+ | 1.2856 | 220 | 0.0708 | - | - | - | - | - |
755
+ | 1.3440 | 230 | 0.0827 | - | - | - | - | - |
756
+ | 1.4025 | 240 | 0.0505 | - | - | - | - | - |
757
+ | 1.4609 | 250 | 0.0468 | - | - | - | - | - |
758
+ | 1.5194 | 260 | 0.0371 | - | - | - | - | - |
759
+ | 1.5778 | 270 | 0.049 | - | - | - | - | - |
760
+ | 1.6362 | 280 | 0.0527 | - | - | - | - | - |
761
+ | 1.6947 | 290 | 0.0316 | - | - | - | - | - |
762
+ | 1.7531 | 300 | 0.052 | - | - | - | - | - |
763
+ | 1.8115 | 310 | 0.0298 | - | - | - | - | - |
764
+ | 1.8700 | 320 | 0.0334 | - | - | - | - | - |
765
+ | 1.9284 | 330 | 0.0431 | - | - | - | - | - |
766
+ | 1.9869 | 340 | 0.0316 | - | - | - | - | - |
767
+ | 1.9985 | 342 | - | 0.8216 | 0.8342 | 0.8397 | 0.8006 | 0.8408 |
768
+ | 2.0453 | 350 | 0.0275 | - | - | - | - | - |
769
+ | 2.1037 | 360 | 0.0461 | - | - | - | - | - |
770
+ | 2.1622 | 370 | 0.0341 | - | - | - | - | - |
771
+ | 2.2206 | 380 | 0.0323 | - | - | - | - | - |
772
+ | 2.2790 | 390 | 0.0205 | - | - | - | - | - |
773
+ | 2.3375 | 400 | 0.0223 | - | - | - | - | - |
774
+ | 2.3959 | 410 | 0.0189 | - | - | - | - | - |
775
+ | 2.4543 | 420 | 0.0181 | - | - | - | - | - |
776
+ | 2.5128 | 430 | 0.0144 | - | - | - | - | - |
777
+ | 2.5712 | 440 | 0.0179 | - | - | - | - | - |
778
+ | 2.6297 | 450 | 0.0217 | - | - | - | - | - |
779
+ | 2.6881 | 460 | 0.016 | - | - | - | - | - |
780
+ | 2.7465 | 470 | 0.0143 | - | - | - | - | - |
781
+ | 2.8050 | 480 | 0.0193 | - | - | - | - | - |
782
+ | 2.8634 | 490 | 0.0183 | - | - | - | - | - |
783
+ | 2.9218 | 500 | 0.0171 | - | - | - | - | - |
784
+ | 2.9803 | 510 | 0.0195 | - | - | - | - | - |
785
+ | 2.9978 | 513 | - | 0.8242 | 0.8350 | 0.8409 | 0.8051 | 0.8413 |
786
+ | 3.0387 | 520 | 0.0127 | - | - | - | - | - |
787
+ | 3.0972 | 530 | 0.0261 | - | - | - | - | - |
788
+ | 3.1556 | 540 | 0.017 | - | - | - | - | - |
789
+ | 3.2140 | 550 | 0.0198 | - | - | - | - | - |
790
+ | 3.2725 | 560 | 0.0131 | - | - | - | - | - |
791
+ | 3.3309 | 570 | 0.0156 | - | - | - | - | - |
792
+ | 3.3893 | 580 | 0.0107 | - | - | - | - | - |
793
+ | 3.4478 | 590 | 0.0123 | - | - | - | - | - |
794
+ | 3.5062 | 600 | 0.0111 | - | - | - | - | - |
795
+ | 3.5646 | 610 | 0.0112 | - | - | - | - | - |
796
+ | 3.6231 | 620 | 0.0143 | - | - | - | - | - |
797
+ | 3.6815 | 630 | 0.013 | - | - | - | - | - |
798
+ | 3.7400 | 640 | 0.0105 | - | - | - | - | - |
799
+ | 3.7984 | 650 | 0.0126 | - | - | - | - | - |
800
+ | 3.8568 | 660 | 0.0118 | - | - | - | - | - |
801
+ | 3.9153 | 670 | 0.0163 | - | - | - | - | - |
802
+ | 3.9737 | 680 | 0.0187 | - | - | - | - | - |
803
+ | **3.9971** | **684** | **-** | **0.8248** | **0.8361** | **0.8405** | **0.8058** | **0.8422** |
804
 
805
  * The bold row denotes the saved checkpoint.
806
 
807
  ### Framework Versions
808
  - Python: 3.12.5
809
+ - Sentence Transformers: 3.1.1
810
  - Transformers: 4.44.2
811
+ - PyTorch: 2.4.1
812
  - Accelerate: 0.33.0
813
  - Datasets: 2.21.0
814
  - Tokenizers: 0.19.1
 
833
  #### MatryoshkaLoss
834
  ```bibtex
835
  @misc{kusupati2024matryoshka,
836
+ title={Matryoshka Representation Learning},
837
  author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
838
  year={2024},
839
  eprint={2205.13147},
 
845
  #### MultipleNegativesRankingLoss
846
  ```bibtex
847
  @misc{henderson2017efficient,
848
+ title={Efficient Natural Language Response Suggestion for Smart Reply},
849
  author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
850
  year={2017},
851
  eprint={1705.00652},
config_sentence_transformers.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "__version__": {
3
- "sentence_transformers": "3.0.1",
4
  "transformers": "4.44.2",
5
- "pytorch": "2.4.0"
6
  },
7
  "prompts": {},
8
  "default_prompt_name": null,
 
1
  {
2
  "__version__": {
3
+ "sentence_transformers": "3.1.1",
4
  "transformers": "4.44.2",
5
+ "pytorch": "2.4.1"
6
  },
7
  "prompts": {},
8
  "default_prompt_name": null,
model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:905ef4cf9bbd19190b586e4dc4173774b01b619eb4f4ab5fa233fb5dfc725472
3
  size 555896347
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50557ba011052514c2e8d048257176d199c9b609b15d5c39dfe8a18383aadb60
3
  size 555896347
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:aa8280e8c799e0e3acd77f0263ef9df6449c601d630fdd734bf422c5f5352daf
3
  size 547119128
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6d13bedb560025363d482eb733dc9f18a0bfe5b342490240b9e9faabccf22983
3
  size 547119128
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e93f7b36897bd6605d5c5db5bd2469f5e524cc3d49234bbcc23e781c26404414
3
  size 5496
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d2ca4f4a8289d13a34ff02bc7fb1d07d6772ee0b83178f7a016e3de1e0ba45e
3
  size 5496