Training in progress, step 1, checkpoint

checkpoint-1/README.md

@@ -168,34 +168,34 @@ model-index:
       type: sts-test
     metrics:
     - type: pearson_cosine
-      value: 0.01688864747186382
+      value: 0.033928485348000664
       name: Pearson Cosine
     - type: spearman_cosine
-      value: 0.0704767781934101
+      value: 0.08944249572062771
       name: Spearman Cosine
     - type: pearson_manhattan
-      value: 0.05915868944955206
+      value: 0.06296467882181725
       name: Pearson Manhattan
     - type: spearman_manhattan
-      value: 0.07640924890718144
+      value: 0.08266825793291849
       name: Spearman Manhattan
     - type: pearson_euclidean
-      value: 0.02427633810266949
+      value: 0.03489200141716902
       name: Pearson Euclidean
     - type: spearman_euclidean
-      value: 0.0506775286593327
+      value: 0.06202473500014035
       name: Spearman Euclidean
     - type: pearson_dot
-      value: 0.16020781347065607
+      value: 0.2554086617921545
       name: Pearson Dot
     - type: spearman_dot
-      value: 0.19413812590183685
+      value: 0.27863958137561534
       name: Spearman Dot
     - type: pearson_max
-      value: 0.16020781347065607
+      value: 0.2554086617921545
       name: Pearson Max
     - type: spearman_max
-      value: 0.19413812590183685
+      value: 0.27863958137561534
       name: Spearman Max
   - task:
       type: triplet
@@ -230,55 +230,55 @@ model-index:
       value: 0.55078125
       name: Cosine Accuracy
     - type: cosine_accuracy_threshold
-      value: 0.9466925859451294
+      value: 0.9503422379493713
       name: Cosine Accuracy Threshold
     - type: cosine_f1
-      value: 0.6525198938992042
+      value: 0.6542553191489362
       name: Cosine F1
     - type: cosine_f1_threshold
-      value: 0.49584439396858215
+      value: 0.656802773475647
       name: Cosine F1 Threshold
     - type: cosine_precision
-      value: 0.484251968503937
+      value: 0.48616600790513836
       name: Cosine Precision
     - type: cosine_recall
       value: 1.0
       name: Cosine Recall
     - type: cosine_ap
-      value: 0.5153192822743842
+      value: 0.5203148129920425
       name: Cosine Ap
     - type: dot_accuracy
       value: 0.55078125
       name: Dot Accuracy
     - type: dot_accuracy_threshold
-      value: 417.46221923828125
+      value: 425.30816650390625
       name: Dot Accuracy Threshold
     - type: dot_f1
-      value: 0.6525198938992042
+      value: 0.6542553191489362
       name: Dot F1
     - type: dot_f1_threshold
-      value: 199.873291015625
+      value: 262.8174743652344
       name: Dot F1 Threshold
     - type: dot_precision
-      value: 0.484251968503937
+      value: 0.48616600790513836
       name: Dot Precision
     - type: dot_recall
       value: 1.0
       name: Dot Recall
     - type: dot_ap
-      value: 0.5127659553715838
+      value: 0.5120444819966403
       name: Dot Ap
     - type: manhattan_accuracy
-      value: 0.546875
+      value: 0.5390625
       name: Manhattan Accuracy
     - type: manhattan_accuracy_threshold
-      value: 117.19680786132812
+      value: 107.76934814453125
       name: Manhattan Accuracy Threshold
     - type: manhattan_f1
       value: 0.6542553191489362
       name: Manhattan F1
     - type: manhattan_f1_threshold
-      value: 292.8346252441406
+      value: 271.5865478515625
       name: Manhattan F1 Threshold
     - type: manhattan_precision
       value: 0.48616600790513836
@@ -287,40 +287,40 @@ model-index:
       value: 1.0
       name: Manhattan Recall
     - type: manhattan_ap
-      value: 0.5178540867523715
+      value: 0.5208015383309144
       name: Manhattan Ap
     - type: euclidean_accuracy
-      value: 0.546875
+      value: 0.55078125
       name: Euclidean Accuracy
     - type: euclidean_accuracy_threshold
-      value: 6.84520149230957
+      value: 7.050784111022949
       name: Euclidean Accuracy Threshold
     - type: euclidean_f1
-      value: 0.6525198938992042
+      value: 0.6507936507936508
       name: Euclidean F1
     - type: euclidean_f1_threshold
-      value: 20.29159164428711
+      value: 17.465972900390625
       name: Euclidean F1 Threshold
     - type: euclidean_precision
-      value: 0.484251968503937
+      value: 0.4823529411764706
       name: Euclidean Precision
     - type: euclidean_recall
       value: 1.0
       name: Euclidean Recall
     - type: euclidean_ap
-      value: 0.5128797056139347
+      value: 0.5175301700973289
       name: Euclidean Ap
     - type: max_accuracy
       value: 0.55078125
       name: Max Accuracy
     - type: max_accuracy_threshold
-      value: 417.46221923828125
+      value: 425.30816650390625
       name: Max Accuracy Threshold
     - type: max_f1
       value: 0.6542553191489362
       name: Max F1
     - type: max_f1_threshold
-      value: 292.8346252441406
+      value: 271.5865478515625
       name: Max F1 Threshold
     - type: max_precision
       value: 0.48616600790513836
@@ -329,7 +329,7 @@ model-index:
       value: 1.0
       name: Max Recall
     - type: max_ap
-      value: 0.5178540867523715
+      value: 0.5208015383309144
       name: Max Ap
 ---
 
@@ -392,7 +392,7 @@ Then you can load this model and run inference.
 from sentence_transformers import SentenceTransformer
 
 # Download from the 🤗 Hub
-model = SentenceTransformer("bobox/DeBERTa-small-ST-v1-toytest-checkpoints-tmp")
+model = SentenceTransformer("bobox/DeBERTa-small-ST-v1-toytest")
 # Run inference
 sentences = [
     'when is season 2 of the ranch coming to netflix',
@@ -443,16 +443,16 @@ You can finetune this model on your own dataset.
 
 | Metric              | Value      |
 |:--------------------|:-----------|
-| pearson_cosine      | 0.0169     |
-| **spearman_cosine** | **0.0705** |
-| pearson_manhattan   | 0.0592     |
-| spearman_manhattan  | 0.0764     |
-| pearson_euclidean   | 0.0243     |
-| spearman_euclidean  | 0.0507     |
-| pearson_dot         | 0.1602     |
-| spearman_dot        | 0.1941     |
-| pearson_max         | 0.1602     |
-| spearman_max        | 0.1941     |
+| pearson_cosine      | 0.0339     |
+| **spearman_cosine** | **0.0894** |
+| pearson_manhattan   | 0.063      |
+| spearman_manhattan  | 0.0827     |
+| pearson_euclidean   | 0.0349     |
+| spearman_euclidean  | 0.062      |
+| pearson_dot         | 0.2554     |
+| spearman_dot        | 0.2786     |
+| pearson_max         | 0.2554     |
+| spearman_max        | 0.2786     |
 
 #### Triplet
 * Dataset: `NLI-v2`
@@ -473,40 +473,40 @@ You can finetune this model on your own dataset.
 | Metric                       | Value      |
 |:-----------------------------|:-----------|
 | cosine_accuracy              | 0.5508     |
-| cosine_accuracy_threshold    | 0.9467     |
-| cosine_f1                    | 0.6525     |
-| cosine_f1_threshold          | 0.4958     |
-| cosine_precision             | 0.4843     |
+| cosine_accuracy_threshold    | 0.9503     |
+| cosine_f1                    | 0.6543     |
+| cosine_f1_threshold          | 0.6568     |
+| cosine_precision             | 0.4862     |
 | cosine_recall                | 1.0        |
-| cosine_ap                    | 0.5153     |
+| cosine_ap                    | 0.5203     |
 | dot_accuracy                 | 0.5508     |
-| dot_accuracy_threshold       | 417.4622   |
-| dot_f1                       | 0.6525     |
-| dot_f1_threshold             | 199.8733   |
-| dot_precision                | 0.4843     |
+| dot_accuracy_threshold       | 425.3082   |
+| dot_f1                       | 0.6543     |
+| dot_f1_threshold             | 262.8175   |
+| dot_precision                | 0.4862     |
 | dot_recall                   | 1.0        |
-| dot_ap                       | 0.5128     |
-| manhattan_accuracy           | 0.5469     |
-| manhattan_accuracy_threshold | 117.1968   |
+| dot_ap                       | 0.512      |
+| manhattan_accuracy           | 0.5391     |
+| manhattan_accuracy_threshold | 107.7693   |
 | manhattan_f1                 | 0.6543     |
-| manhattan_f1_threshold       | 292.8346   |
+| manhattan_f1_threshold       | 271.5865   |
 | manhattan_precision          | 0.4862     |
 | manhattan_recall             | 1.0        |
-| manhattan_ap                 | 0.5179     |
-| euclidean_accuracy           | 0.5469     |
-| euclidean_accuracy_threshold | 6.8452     |
-| euclidean_f1                 | 0.6525     |
-| euclidean_f1_threshold       | 20.2916    |
-| euclidean_precision          | 0.4843     |
+| manhattan_ap                 | 0.5208     |
+| euclidean_accuracy           | 0.5508     |
+| euclidean_accuracy_threshold | 7.0508     |
+| euclidean_f1                 | 0.6508     |
+| euclidean_f1_threshold       | 17.466     |
+| euclidean_precision          | 0.4824     |
 | euclidean_recall             | 1.0        |
-| euclidean_ap                 | 0.5129     |
+| euclidean_ap                 | 0.5175     |
 | max_accuracy                 | 0.5508     |
-| max_accuracy_threshold       | 417.4622   |
+| max_accuracy_threshold       | 425.3082   |
 | max_f1                       | 0.6543     |
-| max_f1_threshold             | 292.8346   |
+| max_f1_threshold             | 271.5865   |
 | max_precision                | 0.4862     |
 | max_recall                   | 1.0        |
-| **max_ap**                   | **0.5179** |
+| **max_ap**                   | **0.5208** |
 
 <!--
 ## Bias, Risks and Limitations
@@ -1155,15 +1155,15 @@ You can finetune this model on your own dataset.
 #### Non-Default Hyperparameters
 
 - `eval_strategy`: steps
-- `per_device_train_batch_size`: 160
+- `per_device_train_batch_size`: 320
 - `per_device_eval_batch_size`: 64
-- `gradient_accumulation_steps`: 8
+- `gradient_accumulation_steps`: 4
 - `learning_rate`: 4e-05
-- `weight_decay`: 0.0001
+- `weight_decay`: 5e-05
 - `num_train_epochs`: 0.1
 - `lr_scheduler_type`: cosine_with_min_lr
-- `lr_scheduler_kwargs`: {'num_cycles': 0.5, 'min_lr': 1.3333333333333335e-05}
-- `warmup_ratio`: 0.33
+- `lr_scheduler_kwargs`: {'num_cycles': 0.5, 'min_lr': 1e-05}
+- `warmup_ratio`: 0.15
 - `save_safetensors`: False
 - `fp16`: True
 - `push_to_hub`: True
@@ -1178,14 +1178,14 @@ You can finetune this model on your own dataset.
 - `do_predict`: False
 - `eval_strategy`: steps
 - `prediction_loss_only`: True
-- `per_device_train_batch_size`: 160
+- `per_device_train_batch_size`: 320
 - `per_device_eval_batch_size`: 64
 - `per_gpu_train_batch_size`: None
 - `per_gpu_eval_batch_size`: None
-- `gradient_accumulation_steps`: 8
+- `gradient_accumulation_steps`: 4
 - `eval_accumulation_steps`: None
 - `learning_rate`: 4e-05
-- `weight_decay`: 0.0001
+- `weight_decay`: 5e-05
 - `adam_beta1`: 0.9
 - `adam_beta2`: 0.999
 - `adam_epsilon`: 1e-08
@@ -1193,8 +1193,8 @@ You can finetune this model on your own dataset.
 - `num_train_epochs`: 0.1
 - `max_steps`: -1
 - `lr_scheduler_type`: cosine_with_min_lr
-- `lr_scheduler_kwargs`: {'num_cycles': 0.5, 'min_lr': 1.3333333333333335e-05}
-- `warmup_ratio`: 0.33
+- `lr_scheduler_kwargs`: {'num_cycles': 0.5, 'min_lr': 1e-05}
+- `warmup_ratio`: 0.15
 - `warmup_steps`: 0
 - `log_level`: passive
 - `log_level_replica`: warning
@@ -1290,6 +1290,8 @@ You can finetune this model on your own dataset.
 | Epoch  | Step | Training Loss | negation-triplets loss | vitaminc-pairs loss | qasc pairs loss | scitail-pairs-pos loss | gooaq pairs loss | xsum-pairs loss | paws-pos loss | nq pairs loss | msmarco pairs loss | openbookqa pairs loss | trivia pairs loss | sciq pairs loss | NLI-v2_max_accuracy | VitaminC_max_ap | sts-test_spearman_cosine |
 |:------:|:----:|:-------------:|:----------------------:|:-------------------:|:---------------:|:----------------------:|:----------------:|:---------------:|:-------------:|:-------------:|:------------------:|:---------------------:|:-----------------:|:---------------:|:-------------------:|:---------------:|:------------------------:|
 | 0.0548 | 1    | 6.851         | 5.2593                 | 2.7279              | 7.9013          | 1.9180                 | 8.1263           | 6.3900          | 2.2178        | 10.4461       | 10.6071            | 4.7477                | 7.8702            | 1.1206          | 1.0                 | 0.5179          | 0.0705                   |
+| 0.1096 | 2    | 7.0772        | 5.2441                 | 2.6973              | 6.5699          | 1.9754                 | 6.6944           | 6.1687          | 2.3460        | 8.0334        | 7.9983             | 4.5152                | 6.7688            | 0.9838          | 1.0                 | 0.5208          | 0.0894                   |
+| 0.0519 | 1    | 7.4907        | 5.2441                 | 2.6973              | 6.5699          | 1.9754                 | 6.6944           | 6.1687          | 2.3460        | 8.0334        | 7.9983             | 4.5152                | 6.7688            | 0.9838          | 1.0                 | 0.5208          | 0.0894                   |
 
 
 ### Framework Versions

checkpoint-1/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:aabe9a3d980e783926dd69a89875e5beb9cc987ee70c92c26e2fb87a94434131
+oid sha256:2559a3b3dfddec8028d1fe2bf75a77ee9151c2eab3dedeb6242d0d710fc8a441
 size 1130520122

checkpoint-1/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:746e90be4f1fd2037efed931baab61aec80c17d7b250b7618290bcd8ec2c09fb
+oid sha256:deabaff50d850f82996bcefd758e8747428f8bf3eb8ff3287402bd7ebb10f5f8
 size 565251810

checkpoint-1/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:39e31a009a68f800d7116d7dc544c153ec26170c60d13a9695fe05d2a6366f41
+oid sha256:d0c73f90426d27a0931c2aab6eaba0594fa6ddaafc6b254dfc4cf59a3d75971a
 size 14244

checkpoint-1/trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.0547945205479452,
+  "epoch": 0.05194805194805195,
   "eval_steps": 1,
   "global_step": 1,
   "is_hyper_param_search": false,
@@ -9,157 +9,157 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.0547945205479452,
-      "grad_norm": 17.049646377563477,
+      "epoch": 0.05194805194805195,
+      "grad_norm": 15.329390525817871,
       "learning_rate": 4e-05,
-      "loss": 6.851,
+      "loss": 7.4907,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
+      "epoch": 0.05194805194805195,
       "eval_NLI-v2_cosine_accuracy": 1.0,
       "eval_NLI-v2_dot_accuracy": 0.125,
       "eval_NLI-v2_euclidean_accuracy": 1.0,
       "eval_NLI-v2_manhattan_accuracy": 1.0,
       "eval_NLI-v2_max_accuracy": 1.0,
       "eval_VitaminC_cosine_accuracy": 0.55078125,
-      "eval_VitaminC_cosine_accuracy_threshold": 0.9466925859451294,
-      "eval_VitaminC_cosine_ap": 0.5153192822743842,
-      "eval_VitaminC_cosine_f1": 0.6525198938992042,
-      "eval_VitaminC_cosine_f1_threshold": 0.49584439396858215,
-      "eval_VitaminC_cosine_precision": 0.484251968503937,
+      "eval_VitaminC_cosine_accuracy_threshold": 0.9503422379493713,
+      "eval_VitaminC_cosine_ap": 0.5203148129920425,
+      "eval_VitaminC_cosine_f1": 0.6542553191489362,
+      "eval_VitaminC_cosine_f1_threshold": 0.656802773475647,
+      "eval_VitaminC_cosine_precision": 0.48616600790513836,
       "eval_VitaminC_cosine_recall": 1.0,
       "eval_VitaminC_dot_accuracy": 0.55078125,
-      "eval_VitaminC_dot_accuracy_threshold": 417.46221923828125,
-      "eval_VitaminC_dot_ap": 0.5127659553715838,
-      "eval_VitaminC_dot_f1": 0.6525198938992042,
-      "eval_VitaminC_dot_f1_threshold": 199.873291015625,
-      "eval_VitaminC_dot_precision": 0.484251968503937,
+      "eval_VitaminC_dot_accuracy_threshold": 425.30816650390625,
+      "eval_VitaminC_dot_ap": 0.5120444819966403,
+      "eval_VitaminC_dot_f1": 0.6542553191489362,
+      "eval_VitaminC_dot_f1_threshold": 262.8174743652344,
+      "eval_VitaminC_dot_precision": 0.48616600790513836,
       "eval_VitaminC_dot_recall": 1.0,
-      "eval_VitaminC_euclidean_accuracy": 0.546875,
-      "eval_VitaminC_euclidean_accuracy_threshold": 6.84520149230957,
-      "eval_VitaminC_euclidean_ap": 0.5128797056139347,
-      "eval_VitaminC_euclidean_f1": 0.6525198938992042,
-      "eval_VitaminC_euclidean_f1_threshold": 20.29159164428711,
-      "eval_VitaminC_euclidean_precision": 0.484251968503937,
+      "eval_VitaminC_euclidean_accuracy": 0.55078125,
+      "eval_VitaminC_euclidean_accuracy_threshold": 7.050784111022949,
+      "eval_VitaminC_euclidean_ap": 0.5175301700973289,
+      "eval_VitaminC_euclidean_f1": 0.6507936507936508,
+      "eval_VitaminC_euclidean_f1_threshold": 17.465972900390625,
+      "eval_VitaminC_euclidean_precision": 0.4823529411764706,
       "eval_VitaminC_euclidean_recall": 1.0,
-      "eval_VitaminC_manhattan_accuracy": 0.546875,
-      "eval_VitaminC_manhattan_accuracy_threshold": 117.19680786132812,
-      "eval_VitaminC_manhattan_ap": 0.5178540867523715,
+      "eval_VitaminC_manhattan_accuracy": 0.5390625,
+      "eval_VitaminC_manhattan_accuracy_threshold": 107.76934814453125,
+      "eval_VitaminC_manhattan_ap": 0.5208015383309144,
       "eval_VitaminC_manhattan_f1": 0.6542553191489362,
-      "eval_VitaminC_manhattan_f1_threshold": 292.8346252441406,
+      "eval_VitaminC_manhattan_f1_threshold": 271.5865478515625,
       "eval_VitaminC_manhattan_precision": 0.48616600790513836,
       "eval_VitaminC_manhattan_recall": 1.0,
       "eval_VitaminC_max_accuracy": 0.55078125,
-      "eval_VitaminC_max_accuracy_threshold": 417.46221923828125,
-      "eval_VitaminC_max_ap": 0.5178540867523715,
+      "eval_VitaminC_max_accuracy_threshold": 425.30816650390625,
+      "eval_VitaminC_max_ap": 0.5208015383309144,
       "eval_VitaminC_max_f1": 0.6542553191489362,
-      "eval_VitaminC_max_f1_threshold": 292.8346252441406,
+      "eval_VitaminC_max_f1_threshold": 271.5865478515625,
       "eval_VitaminC_max_precision": 0.48616600790513836,
       "eval_VitaminC_max_recall": 1.0,
-      "eval_sequential_score": 0.5178540867523715,
-      "eval_sts-test_pearson_cosine": 0.01688864747186382,
-      "eval_sts-test_pearson_dot": 0.16020781347065607,
-      "eval_sts-test_pearson_euclidean": 0.02427633810266949,
-      "eval_sts-test_pearson_manhattan": 0.05915868944955206,
-      "eval_sts-test_pearson_max": 0.16020781347065607,
-      "eval_sts-test_spearman_cosine": 0.0704767781934101,
-      "eval_sts-test_spearman_dot": 0.19413812590183685,
-      "eval_sts-test_spearman_euclidean": 0.0506775286593327,
-      "eval_sts-test_spearman_manhattan": 0.07640924890718144,
-      "eval_sts-test_spearman_max": 0.19413812590183685,
-      "eval_vitaminc-pairs_loss": 2.7278947830200195,
-      "eval_vitaminc-pairs_runtime": 1.5228,
-      "eval_vitaminc-pairs_samples_per_second": 70.92,
-      "eval_vitaminc-pairs_steps_per_second": 1.313,
+      "eval_sequential_score": 0.5208015383309144,
+      "eval_sts-test_pearson_cosine": 0.033928485348000664,
+      "eval_sts-test_pearson_dot": 0.2554086617921545,
+      "eval_sts-test_pearson_euclidean": 0.03489200141716902,
+      "eval_sts-test_pearson_manhattan": 0.06296467882181725,
+      "eval_sts-test_pearson_max": 0.2554086617921545,
+      "eval_sts-test_spearman_cosine": 0.08944249572062771,
+      "eval_sts-test_spearman_dot": 0.27863958137561534,
+      "eval_sts-test_spearman_euclidean": 0.06202473500014035,
+      "eval_sts-test_spearman_manhattan": 0.08266825793291849,
+      "eval_sts-test_spearman_max": 0.27863958137561534,
+      "eval_vitaminc-pairs_loss": 2.6973235607147217,
+      "eval_vitaminc-pairs_runtime": 1.4352,
+      "eval_vitaminc-pairs_samples_per_second": 75.251,
+      "eval_vitaminc-pairs_steps_per_second": 1.394,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_negation-triplets_loss": 5.259257793426514,
-      "eval_negation-triplets_runtime": 0.3007,
-      "eval_negation-triplets_samples_per_second": 212.85,
-      "eval_negation-triplets_steps_per_second": 3.326,
+      "epoch": 0.05194805194805195,
+      "eval_negation-triplets_loss": 5.244063854217529,
+      "eval_negation-triplets_runtime": 0.293,
+      "eval_negation-triplets_samples_per_second": 218.403,
+      "eval_negation-triplets_steps_per_second": 3.413,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_scitail-pairs-pos_loss": 1.9180195331573486,
-      "eval_scitail-pairs-pos_runtime": 0.4022,
-      "eval_scitail-pairs-pos_samples_per_second": 134.25,
-      "eval_scitail-pairs-pos_steps_per_second": 2.486,
+      "epoch": 0.05194805194805195,
+      "eval_scitail-pairs-pos_loss": 1.975368618965149,
+      "eval_scitail-pairs-pos_runtime": 0.3648,
+      "eval_scitail-pairs-pos_samples_per_second": 148.035,
+      "eval_scitail-pairs-pos_steps_per_second": 2.741,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_xsum-pairs_loss": 6.389987945556641,
-      "eval_xsum-pairs_runtime": 3.3036,
-      "eval_xsum-pairs_samples_per_second": 38.746,
-      "eval_xsum-pairs_steps_per_second": 0.605,
+      "epoch": 0.05194805194805195,
+      "eval_xsum-pairs_loss": 6.168678283691406,
+      "eval_xsum-pairs_runtime": 3.3419,
+      "eval_xsum-pairs_samples_per_second": 38.302,
+      "eval_xsum-pairs_steps_per_second": 0.598,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_sciq_pairs_loss": 1.1206029653549194,
-      "eval_sciq_pairs_runtime": 3.3516,
-      "eval_sciq_pairs_samples_per_second": 38.191,
-      "eval_sciq_pairs_steps_per_second": 0.597,
+      "epoch": 0.05194805194805195,
+      "eval_sciq_pairs_loss": 0.9838346838951111,
+      "eval_sciq_pairs_runtime": 3.3684,
+      "eval_sciq_pairs_samples_per_second": 38.0,
+      "eval_sciq_pairs_steps_per_second": 0.594,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_qasc_pairs_loss": 7.901282787322998,
-      "eval_qasc_pairs_runtime": 0.6747,
-      "eval_qasc_pairs_samples_per_second": 189.709,
+      "epoch": 0.05194805194805195,
+      "eval_qasc_pairs_loss": 6.569870471954346,
+      "eval_qasc_pairs_runtime": 0.6748,
+      "eval_qasc_pairs_samples_per_second": 189.699,
       "eval_qasc_pairs_steps_per_second": 2.964,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_openbookqa_pairs_loss": 4.747707366943359,
-      "eval_openbookqa_pairs_runtime": 0.5849,
-      "eval_openbookqa_pairs_samples_per_second": 218.847,
-      "eval_openbookqa_pairs_steps_per_second": 3.419,
+      "epoch": 0.05194805194805195,
+      "eval_openbookqa_pairs_loss": 4.515223979949951,
+      "eval_openbookqa_pairs_runtime": 0.5729,
+      "eval_openbookqa_pairs_samples_per_second": 223.438,
+      "eval_openbookqa_pairs_steps_per_second": 3.491,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_msmarco_pairs_loss": 10.60707950592041,
-      "eval_msmarco_pairs_runtime": 1.2013,
-      "eval_msmarco_pairs_samples_per_second": 106.551,
-      "eval_msmarco_pairs_steps_per_second": 1.665,
+      "epoch": 0.05194805194805195,
+      "eval_msmarco_pairs_loss": 7.998310565948486,
+      "eval_msmarco_pairs_runtime": 1.2044,
+      "eval_msmarco_pairs_samples_per_second": 106.278,
+      "eval_msmarco_pairs_steps_per_second": 1.661,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_nq_pairs_loss": 10.446059226989746,
-      "eval_nq_pairs_runtime": 2.7595,
-      "eval_nq_pairs_samples_per_second": 46.385,
-      "eval_nq_pairs_steps_per_second": 0.725,
+      "epoch": 0.05194805194805195,
+      "eval_nq_pairs_loss": 8.03341007232666,
+      "eval_nq_pairs_runtime": 2.7885,
+      "eval_nq_pairs_samples_per_second": 45.903,
+      "eval_nq_pairs_steps_per_second": 0.717,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_trivia_pairs_loss": 7.870224952697754,
-      "eval_trivia_pairs_runtime": 3.9907,
-      "eval_trivia_pairs_samples_per_second": 29.819,
-      "eval_trivia_pairs_steps_per_second": 0.501,
+      "epoch": 0.05194805194805195,
+      "eval_trivia_pairs_loss": 6.768779754638672,
+      "eval_trivia_pairs_runtime": 4.0886,
+      "eval_trivia_pairs_samples_per_second": 29.105,
+      "eval_trivia_pairs_steps_per_second": 0.489,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_gooaq_pairs_loss": 8.126296997070312,
-      "eval_gooaq_pairs_runtime": 0.8871,
-      "eval_gooaq_pairs_samples_per_second": 144.296,
-      "eval_gooaq_pairs_steps_per_second": 2.255,
+      "epoch": 0.05194805194805195,
+      "eval_gooaq_pairs_loss": 6.694434642791748,
+      "eval_gooaq_pairs_runtime": 0.9221,
+      "eval_gooaq_pairs_samples_per_second": 138.819,
+      "eval_gooaq_pairs_steps_per_second": 2.169,
       "step": 1
     },
     {
-      "epoch": 0.0547945205479452,
-      "eval_paws-pos_loss": 2.2177822589874268,
-      "eval_paws-pos_runtime": 0.69,
-      "eval_paws-pos_samples_per_second": 185.509,
-      "eval_paws-pos_steps_per_second": 2.899,
+      "epoch": 0.05194805194805195,
+      "eval_paws-pos_loss": 2.3459720611572266,
+      "eval_paws-pos_runtime": 0.7324,
+      "eval_paws-pos_samples_per_second": 174.765,
+      "eval_paws-pos_steps_per_second": 2.731,
       "step": 1
     }
   ],
@@ -181,7 +181,7 @@
     }
   },
   "total_flos": 0.0,
-  "train_batch_size": 160,
+  "train_batch_size": 320,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-1/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7d8ee34c18a400f97b88a2eac5f9619f6acd81ed3e37e87a781d225dd83c9b5d
+oid sha256:ce662bf5dcaba289d862613e794cf1abf75d3964b4ccae6eef4c74365a1d454b
 size 5688