Hugging Face's logo

newbie-geek
/
tinyllama-v1-finetune

PEFT
Safetensors
Generated from Trainer
Model card Files Community
tinyllama-v1-finetune / README.md
newbie-geek's picture
newbie-geek
End of training
38121cb verified
preview code
|
raw
history blame
5.51 kB
metadata
license: apache-2.0
library_name: peft
tags:
  - generated_from_trainer
base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
model-index:
  - name: tinyllama-v1-finetune
    results: []

tinyllama-v1-finetune

This model is a fine-tuned version of TinyLlama/TinyLlama-1.1B-Chat-v1.0 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5205

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 2.5e-05
  • train_batch_size: 2
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • training_steps: 400
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
2.7664 0.04 5 2.7807
2.7326 0.08 10 2.7249
2.6955 0.12 15 2.6739
2.643 0.16 20 2.6085
2.6056 0.2 25 2.5554
2.5018 0.24 30 2.4923
2.4587 0.28 35 2.4204
2.3997 0.32 40 2.3484
2.3142 0.36 45 2.2673
2.2507 0.4 50 2.1903
2.1405 0.43 55 2.1071
2.0471 0.47 60 2.0212
1.9957 0.51 65 1.9385
1.9007 0.55 70 1.8448
1.8046 0.59 75 1.7501
1.691 0.63 80 1.6495
1.5957 0.67 85 1.5484
1.53 0.71 90 1.4608
1.4161 0.75 95 1.3669
1.3409 0.79 100 1.2772
1.239 0.83 105 1.1833
1.1591 0.87 110 1.1122
1.0729 0.91 115 1.0276
1.0237 0.95 120 0.9615
0.9293 0.99 125 0.9021
0.891 1.03 130 0.8531
0.8365 1.07 135 0.8155
0.7876 1.11 140 0.7898
0.7821 1.15 145 0.7669
0.7392 1.19 150 0.7516
0.77 1.23 155 0.7397
0.7088 1.26 160 0.7226
0.7246 1.3 165 0.7101
0.7007 1.34 170 0.6960
0.6667 1.38 175 0.6797
0.6898 1.42 180 0.6666
0.6608 1.46 185 0.6599
0.6526 1.5 190 0.6451
0.6078 1.54 195 0.6350
0.6336 1.58 200 0.6248
0.6074 1.62 205 0.6167
0.6114 1.66 210 0.6131
0.575 1.7 215 0.6041
0.5933 1.74 220 0.5981
0.5983 1.78 225 0.5910
0.5907 1.82 230 0.5845
0.5853 1.86 235 0.5801
0.5881 1.9 240 0.5749
0.5613 1.94 245 0.5700
0.5852 1.98 250 0.5696
0.5781 2.02 255 0.5652
0.5812 2.06 260 0.5609
0.5677 2.09 265 0.5576
0.5544 2.13 270 0.5541
0.536 2.17 275 0.5504
0.5283 2.21 280 0.5487
0.5326 2.25 285 0.5454
0.568 2.29 290 0.5402
0.5448 2.33 295 0.5395
0.5581 2.37 300 0.5377
0.5406 2.41 305 0.5355
0.4996 2.45 310 0.5333
0.5243 2.49 315 0.5346
0.5591 2.53 320 0.5312
0.5122 2.57 325 0.5297
0.5426 2.61 330 0.5290
0.4955 2.65 335 0.5290
0.5531 2.69 340 0.5273
0.5147 2.73 345 0.5278
0.5195 2.77 350 0.5246
0.5268 2.81 355 0.5254
0.5284 2.85 360 0.5236
0.5272 2.89 365 0.5224
0.5053 2.92 370 0.5240
0.528 2.96 375 0.5200
0.517 3.0 380 0.5204
0.5409 3.04 385 0.5206
0.5204 3.08 390 0.5192
0.52 3.12 395 0.5200
0.5142 3.16 400 0.5205

Framework versions

  • PEFT 0.8.1
  • Transformers 4.37.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.1