Edit model card
YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

Quantization made by Richard Erkhov.

Github

Discord

Request more models

TinyLLaMA-1.1B-OrcaPlatty - GGUF

Original model description:

license: apache-2.0 base_model: jeff31415/TinyLlama-1.1B-1T-OpenOrca tags: - generated_from_trainer model-index: - name: results results: []

results

This model is a fine-tuned version of jeff31415/TinyLlama-1.1B-1T-OpenOrca on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.5156

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: 9e-07
  • train_batch_size: 20
  • eval_batch_size: 4
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 80
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss
2.1726 0.03 8 2.3170
2.1444 0.05 16 2.2937
2.1036 0.08 24 2.2707
2.0703 0.1 32 2.2478
2.0604 0.13 40 2.2248
2.046 0.15 48 2.2013
1.9919 0.18 56 2.1780
1.9842 0.21 64 2.1547
1.9234 0.23 72 2.1320
1.9235 0.26 80 2.1099
1.9096 0.28 88 2.0884
1.8722 0.31 96 2.0679
1.8594 0.34 104 2.0479
1.8438 0.36 112 2.0283
1.7581 0.39 120 2.0089
1.7852 0.41 128 1.9901
1.7634 0.44 136 1.9714
1.7296 0.46 144 1.9531
1.6976 0.49 152 1.9353
1.6861 0.52 160 1.9173
1.6683 0.54 168 1.8993
1.6255 0.57 176 1.8826
1.619 0.59 184 1.8673
1.6455 0.62 192 1.8534
1.5784 0.65 200 1.8399
1.6078 0.67 208 1.8259
1.5703 0.7 216 1.8124
1.5215 0.72 224 1.7989
1.542 0.75 232 1.7852
1.5147 0.77 240 1.7721
1.5092 0.8 248 1.7589
1.4564 0.83 256 1.7456
1.4985 0.85 264 1.7324
1.4505 0.88 272 1.7189
1.4447 0.9 280 1.7052
1.4436 0.93 288 1.6924
1.4132 0.95 296 1.6799
1.3791 0.98 304 1.6680
1.3877 1.01 312 1.6565
1.3807 1.03 320 1.6453
1.3391 1.06 328 1.6352
1.3232 1.08 336 1.6251
1.3293 1.11 344 1.6159
1.3029 1.14 352 1.6074
1.3173 1.16 360 1.5992
1.3006 1.19 368 1.5926
1.2547 1.21 376 1.5863
1.2704 1.24 384 1.5805
1.2964 1.26 392 1.5749
1.277 1.29 400 1.5695
1.2718 1.32 408 1.5657
1.2379 1.34 416 1.5619
1.2746 1.37 424 1.5585
1.2349 1.39 432 1.5559
1.2264 1.42 440 1.5531
1.2365 1.45 448 1.5505
1.2242 1.47 456 1.5484
1.2094 1.5 464 1.5462
1.2196 1.52 472 1.5444
1.2447 1.55 480 1.5426
1.2127 1.57 488 1.5407
1.2278 1.6 496 1.5391
1.2089 1.63 504 1.5377
1.2069 1.65 512 1.5361
1.2264 1.68 520 1.5350
1.2027 1.7 528 1.5338
1.2138 1.73 536 1.5325
1.207 1.75 544 1.5313
1.2155 1.78 552 1.5304
1.2192 1.81 560 1.5295
1.2223 1.83 568 1.5287
1.2281 1.86 576 1.5278
1.1977 1.88 584 1.5269
1.2101 1.91 592 1.5261
1.2099 1.94 600 1.5254
1.1873 1.96 608 1.5245
1.204 1.99 616 1.5242
1.21 2.01 624 1.5239
1.242 2.04 632 1.5231
1.1696 2.06 640 1.5224
1.1803 2.09 648 1.5218
1.1692 2.12 656 1.5213
1.212 2.14 664 1.5208
1.1977 2.17 672 1.5204
1.187 2.19 680 1.5201
1.1858 2.22 688 1.5199
1.1824 2.25 696 1.5194
1.1914 2.27 704 1.5190
1.1815 2.3 712 1.5187
1.2021 2.32 720 1.5184
1.1872 2.35 728 1.5181
1.1901 2.37 736 1.5178
1.1933 2.4 744 1.5177
1.1773 2.43 752 1.5175
1.1935 2.45 760 1.5172
1.2118 2.48 768 1.5170
1.1816 2.5 776 1.5169
1.1842 2.53 784 1.5167
1.1891 2.55 792 1.5165
1.1883 2.58 800 1.5164
1.1506 2.61 808 1.5163
1.1708 2.63 816 1.5162
1.1944 2.66 824 1.5160
1.1575 2.68 832 1.5159
1.1698 2.71 840 1.5160
1.1525 2.74 848 1.5158
1.1767 2.76 856 1.5157
1.1943 2.79 864 1.5158
1.1727 2.81 872 1.5157
1.195 2.84 880 1.5157
1.1771 2.86 888 1.5157
1.1731 2.89 896 1.5156
1.191 2.92 904 1.5157
1.1903 2.94 912 1.5156
1.1821 2.97 920 1.5156
1.2 2.99 928 1.5156

Framework versions

  • Transformers 4.36.0.dev0
  • Pytorch 2.1.0+cu118
  • Datasets 2.15.1.dev0
  • Tokenizers 0.15.0
Downloads last month
222
GGUF
Model size
1.1B params
Architecture
llama

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

Inference API
Unable to determine this model's library. Check the docs .