Hugging Face's logo

flax-community
/
roberta-base-thai

Fill-Mask
Transformers
PyTorch
TensorBoard
Safetensors
roberta
Model card Files
xet
 Metrics Community
1
roberta-base-thai

Commit History

  1. Adding `safetensors` variant of this model (#1)⁠

    yhavinga +1: SFconvertbot
    Verified signature
    ee83601

  1. update tfevent

    sakares
    3d400e8

  2. update event

    sakares
    596725d

  3. fix run.sh

    sakares
    6665f93

  4. new event

    sakares
    6e833e6

  1. use proper tokenizer.json and change to batch_size=256

    sakares
    de940b9

  1. train with 10K batch_size

    sakares
    89c14b0

  2. remove old tokenizer.json. it's not trained properly

    sakares
    a615d86

  3. change config and fix training tokenizer script

    sakares
    13d02d7

  4. remove tfevents

    sakares
    70b401c

  5. rename previous tfevents

    sakares
    d12a293

  6. add 4th tfevents

    sakares
    8dc7108

  7. increaaase bs=128 and change dtype to fp32

    sakares
    2071335

  8. remove old model

    sakares
    8887f4f

  9. change to proper vocab_size

    sakares
    9fa8adc

  10. update pt model

    sakares
    d6a778b

  11. add pt model from 15K trained steps

    sakares
    1aa50f9

  12. update rule

    sakares
    5a8de81

  13. update evnt

    sakares
    24b543a

  14. update event

    sakares
    578dba5

  1. update event

    sakares
    2c966a4

  2. Saving weights and logs of step 15000

    sakares
    8664a01

  3. check event

    sakares
    d74794e

  1. Saving weights and logs of step 10000

    sakares
    908d1a5

  2. update event

    sakares
    be4ec62

  3. update tfevent

    sakares
    e60615a

  1. Saving weights and logs of step 5000

    sakares
    acba88d

  2. update event

    sakares
    14c6643

  3. Create Thai Tokenizer script

    sakares
    5ad5b72

  4. check event

    sakares
    e76044d

  5. init tfevents for experiment#3

    sakares
    bbeb147

  6. remove old exp&model and lower steps to save

    sakares
    64934c7

  7. fix script

    sakares
    9c0b789

  8. make it resumable training

    sakares
    51e884f

  9. Saving weights and logs of step 10000

    sakares
    3040f73

  1. update event

    sakares
    a9d2138

  2. update tfevent

    sakares
    f8a3350

  1. update event

    sakares
    bae02ca

  2. remove old event

    sakares
    23b10e3

  3. add experiment_2 tfevents

    sakares
    e0c32ef

  4. update script

    sakares
    8976790

  5. update script regarding to https://github.com/huggingface/transformers/pull/12608

    sakares
    6aba5c0

  6. lower lr regarding explode gradient. also new seed number

    sakares
    b0729e9

  7. Merge branch 'main' of https://huggingface.co/flax-community/roberta-base-thai into main

    sakares
    bfd5d73

  8. update tfevents

    sakares
    0d2b32a

  1. Update .gitattributes

    patrickvonplaten
    1e02891

  2. update event

    sakares
    d9006b0

  3. update event

    sakares
    33260f5

  1. update event

    sakares
    0b153f6

  2. update tfevents

    sakares
    e0cfb09