README.md

metadata

license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - emotion
metrics:
  - accuracy
model-index:
  - name: xtremedistil-emotion
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: emotion
          type: emotion
          args: default
        metrics:
          - type: accuracy
            value: 0.9265
            name: Accuracy
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: emotion
          type: emotion
          config: default
          split: test
        metrics:
          - type: accuracy
            value: 0.926
            name: Accuracy
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiYzE3NDg5Y2ZkMDE5OTJmNjYwMTU1MDMwOTUwNTdkOWQ0MWNiZDYxYzUwNDBmNGVkOWU0OWE1MzRiNDYyZDI3NyIsInZlcnNpb24iOjF9.BaDj-FQ6g0cRk7n2MlN2YCb8Iv2VIM2wMwnJeeCTjG15b7TRRfZVtM3CM2WvHymahppscpiqgqPxT7JqkVXkAQ
          - type: precision
            value: 0.8855308537052737
            name: Precision Macro
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiZGQ3MDlmOTdmZTY3Mjc5MmE1ZmFlZTVhOWIxYjA3ZDRmNjM4YmYzNTVmZTYwNmI2OTRmYmE3NDMyOTIxM2RjOSIsInZlcnNpb24iOjF9.r1_TDJRi4RJfhVlFDe83mRtdhqt5KMtvran6qjzRrcwXqNz7prkocFmgNnntn-fqgg6AXgyi6lwVDcuj5L5VBA
          - type: precision
            value: 0.926
            name: Precision Micro
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNzMzMzc4MWY1M2E5Y2M2ZTRiYTc2YzA5YzI4ZWM5MjgzMDgyNjZkMTVjZDYxZGJiMjI0NDdiMWU3ZWM5MjhjYSIsInZlcnNpb24iOjF9.741rqCRY5S8z_QodJ0PvcnccCN79fCE-MeNTEWFegI0oReneULyNOKRulxwxzwY5SN6ILm52xW7km5WJyt8MCg
          - type: precision
            value: 0.9281282413639949
            name: Precision Weighted
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiODVlOTM3ODVhMWM0MjU4Mzg2OGNkYjc2ZmExODYzOWIzYjdlYzE4OWE0ZWI4ZjcxMjJiMGJiMzdhN2RiNTdlNiIsInZlcnNpb24iOjF9.8-HhpgKNt3nTcblnes4KxzsD7Xot3C6Rldp4463H9gaUNBxHcH19mFcpaSaDT_L3mYqetcW891jyNrHoATzuAg
          - type: recall
            value: 0.8969894921856228
            name: Recall Macro
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiYTkxYzZiMzY5YjA3ZjExYmNlNGI4N2Q5NTg0MTcxODgxOTc0MjdhM2FjODAzNjhiNDBjMWY2NWUyMjhhYjNiNSIsInZlcnNpb24iOjF9.t5YyyNtkbaGfLVbFIO15wh6o6BqBIXGTEBheffPax61-cZM0HRQg9BufcHFdZ4dvPd_V_AYWrXdarEm-gLSBBg
          - type: recall
            value: 0.926
            name: Recall Micro
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiZjAxMTUzMmI1YmMwYTBmYzFmM2E3Y2NiY2M4Njc4ZDc1ZWRhMTMyMDVhMWNiMGQ1ZDRiMjcwYmQ0MDAxZmI3NSIsInZlcnNpb24iOjF9.OphK_nR4EkaAUGMdZDq1rP_oBivfLHQhE7XY1HP9izhDd6rV5KobTrSdoxVCHGUtjOm1M6eZqI_1rPpunoCqDQ
          - type: recall
            value: 0.926
            name: Recall Weighted
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMGYxYWZlZmY1MWE4ZTU5YzlmZjA3MjVkZGFlMjk4NjFmMTIwZTNlMWU2ZWE1YWE3ZTc3MzI4NmJhYjM5Y2M5NCIsInZlcnNpb24iOjF9.zRx5GUnSb-T6E3s3NsWn1c1szm63jlB8XeqBUZ3J0m5H6P-QAPcVTaMVn8id-_IExS4g856-dT9YMq3pRh91DQ
          - type: f1
            value: 0.8903400738742536
            name: F1 Macro
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMzE1NDYxYTdiNjAwYzllZmY4ODc1ZTc1YjMyZjA4Njc1NDhjNDM5ZWNmOThjNzQ1MDE5ZDEyMTY0YTljZDcyMiIsInZlcnNpb24iOjF9.j4U3aOySF94GUF94YGA7DPjynVJ7wStBPu8uinEz_AjQFISv8YvHZOO--Kv2S4iKJPQNSGjmqP8jwtVEKt6-AA
          - type: f1
            value: 0.926
            name: F1 Micro
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNTFmYzdiM2FmZDIyMjkxZDk2NGFkMjU4OWJjYzQ1MTJkZThiMmMzYTUzZmJlNjNmYTFlOTRkMTZjODI2NDdiYyIsInZlcnNpb24iOjF9.VY3hvPQL588GY4j9cCJRj1GWZWsdgkRV1F5DKhckC74-w2qFK10zgqSEbb_uhOg3IYLcXev9f8dhIOVcOCPvDg
          - type: f1
            value: 0.9265018282649476
            name: F1 Weighted
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiM2MyNjM2OGMzYzg5ODFiOWI0ZTkxMDAxYTRkNDYwZWIyZGUyYzhhYTUwYWM4NzJhYTk3MGU2N2E5ZTcyNWExMyIsInZlcnNpb24iOjF9.p_7UeUdm-Qy6yfUlZA9EmtAKUzxhfkDTUMkzNRLJ3HD3aFHHwOo8jIY3lEZ-QkucT-jhofgbnQ-jR56HmB1JDw
          - type: loss
            value: 0.2258329838514328
            name: loss
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiZTQwM2Y4NGI0MmQwMDkxMTBiYTdlYjkwNjdiMjVhMGZhOTk0Y2MwMmVlODg2YTczNzg1MGZiMDM2NzIyMzE5ZCIsInZlcnNpb24iOjF9.gCzWQrRm8UsOEcZvT_zC568FZmIcQf8G177IDQmxGVGg1vrOonfnPLX1_xlbcID4vDGeVuw5xYEpxXOAc19GDw

xtremedistil-emotion

This model is a fine-tuned version of microsoft/xtremedistil-l6-h256-uncased on the emotion dataset. It achieves the following results on the evaluation set:

• Accuracy: 0.9265

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 3e-05
• train_batch_size: 128
• eval_batch_size: 8
• seed: 42
• num_epochs: 24

Training results

Epoch	Training Loss	Validation Loss	Accuracy
1	No log	1.238589	0.609000
2	No log	0.934423	0.714000
3	No log	0.768701	0.742000
4	1.074800	0.638208	0.805500
5	1.074800	0.551363	0.851500
6	1.074800	0.476291	0.875500
7	1.074800	0.427313	0.883500
8	0.531500	0.392633	0.886000
9	0.531500	0.357979	0.892000
10	0.531500	0.330304	0.899500
11	0.531500	0.304529	0.907000
12	0.337200	0.287447	0.918000
13	0.337200	0.277067	0.921000
14	0.337200	0.259483	0.921000
15	0.337200	0.257564	0.916500
16	0.246200	0.241970	0.919500
17	0.246200	0.241537	0.921500
18	0.246200	0.235705	0.924500
19	0.246200	0.237325	0.920500
20	0.201400	0.229699	0.923500
21	0.201400	0.227426	0.923000
22	0.201400	0.228554	0.924000
23	0.201400	0.226941	0.925500
24	0.184300	0.225816	0.926500