You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this model content.

SNAC-Denoiser-LLaMA-500M-snac_v2_test_1gpu

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 7.5450

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: 1e-05
  • train_batch_size: 2
  • eval_batch_size: 1
  • seed: 42
  • gradient_accumulation_steps: 16
  • total_train_batch_size: 32
  • optimizer: Use adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 62
  • training_steps: 3124

Training results

Training Loss Epoch Step Validation Loss
9.3455 0.0160 50 9.3244
9.2001 0.0320 100 9.1875
9.1377 0.0480 150 9.1356
9.1092 0.0640 200 9.0910
9.0062 0.0800 250 8.9681
8.8554 0.0960 300 8.8329
8.718 0.1120 350 8.6753
8.5591 0.1280 400 8.5258
8.4367 0.1440 450 8.4062
8.349 0.1600 500 8.3305
8.2861 0.1760 550 8.2693
8.2332 0.1920 600 8.2187
8.2062 0.2080 650 8.1720
8.1408 0.2240 700 8.1226
8.1249 0.2400 750 8.0816
8.0804 0.2560 800 8.0395
8.0492 0.2720 850 8.0060
8.0164 0.2881 900 7.9712
7.9843 0.3041 950 7.9449
7.9496 0.3201 1000 7.9181
7.9486 0.3361 1050 7.8922
7.9317 0.3521 1100 7.8683
7.913 0.3681 1150 7.8490
7.8754 0.3841 1200 7.8260
7.8618 0.4001 1250 7.8044
7.8301 0.4161 1300 7.7877
7.7919 0.4321 1350 7.7684
7.7967 0.4481 1400 7.7496
7.7759 0.4641 1450 7.7350
7.7685 0.4801 1500 7.7192
7.7523 0.4961 1550 7.7041
7.7205 0.5121 1600 7.6902
7.7153 0.5281 1650 7.6767
7.7194 0.5441 1700 7.6648
7.702 0.5601 1750 7.6552
7.7038 0.5761 1800 7.6431
7.694 0.5921 1850 7.6317
7.6717 0.6081 1900 7.6254
7.6509 0.6241 1950 7.6156
7.6552 0.6401 2000 7.6098
7.669 0.6561 2050 7.6022
7.663 0.6721 2100 7.5952
7.6476 0.6881 2150 7.5876
7.6415 0.7041 2200 7.5823
7.6386 0.7201 2250 7.5776
7.6233 0.7361 2300 7.5731
7.6342 0.7521 2350 7.5678
7.6028 0.7681 2400 7.5634
7.6125 0.7841 2450 7.5607
7.6175 0.8001 2500 7.5571
7.6081 0.8161 2550 7.5561
7.6117 0.8321 2600 7.5529
7.5922 0.8482 2650 7.5512
7.6261 0.8642 2700 7.5498
7.5985 0.8802 2750 7.5485
7.6093 0.8962 2800 7.5474
7.6015 0.9122 2850 7.5466
7.5797 0.9282 2900 7.5460
7.621 0.9442 2950 7.5456
7.6041 0.9602 3000 7.5452
7.5733 0.9762 3050 7.5451
7.613 0.9922 3100 7.5450

Framework versions

  • Transformers 4.56.1
  • Pytorch 2.8.0+cu128
  • Datasets 4.0.0
  • Tokenizers 0.22.0
Downloads last month
11
Safetensors
Model size
0.5B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Evaluation results