Edit model card

LLM_Teached_Pegasus_100k

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

  • Loss: 1.5004
  • Rouge1: 0.4923
  • Rouge2: 0.2429
  • Rougel: 0.4134
  • Rougelsum: 0.4134
  • Gen Len: 25.1335
  • Precision: 0.9143
  • Recall: 0.9124
  • F1: 0.9132

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: 2e-05
  • train_batch_size: 32
  • eval_batch_size: 16
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 16
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step F1 Gen Len Validation Loss Precision Recall Rouge1 Rouge2 Rougel Rougelsum
2.1501 1.0 781 0.9072 25.4655 1.7062 0.9082 0.9065 0.4566 0.209 0.3745 0.3744
1.7722 2.0 1562 0.9097 25.4298 1.6314 0.9107 0.909 0.4712 0.2226 0.3906 0.3904
1.7218 3.0 2343 0.9106 25.6569 1.5948 0.9112 0.9103 0.4776 0.2284 0.3965 0.3963
1.6668 4.0 3125 0.9112 25.3451 1.5708 0.9122 0.9107 0.481 0.2316 0.4002 0.4
1.6437 5.0 3906 0.9118 25.482 1.5565 0.9127 0.9113 0.4844 0.2346 0.4034 0.4031
1.6186 6.0 4687 0.912 25.4191 1.5476 0.9129 0.9115 0.4852 0.236 0.4047 0.4044
1.607 7.0 5468 0.9122 25.4949 1.5426 0.9129 0.9118 0.486 0.2367 0.4052 0.405
1.5972 8.0 6248 1.5380 0.4872 0.2387 0.407 0.4071 25.3836 0.9131 0.9118 0.9123
1.5836 9.0 7029 1.5273 0.4891 0.2399 0.4088 0.4089 25.4995 0.9133 0.9122 0.9126
1.5667 10.0 7810 1.5196 0.4906 0.2416 0.411 0.4112 25.3867 0.9135 0.9123 0.9127
1.5521 11.0 8592 1.5124 0.4899 0.2406 0.4102 0.4103 25.2191 0.9137 0.912 0.9127
1.5413 12.0 9373 1.5083 0.4914 0.2416 0.4118 0.412 25.3491 0.9137 0.9123 0.9128
1.5291 13.0 10154 1.5044 0.4913 0.2419 0.4118 0.4119 25.2082 0.914 0.9123 0.913
1.527 14.0 10935 1.5026 0.4917 0.2426 0.4126 0.4128 25.1069 0.9141 0.9123 0.913
1.5203 15.0 11717 1.5006 0.4921 0.243 0.4135 0.4136 25.1062 0.9143 0.9123 0.9131
1.5126 16.0 12496 1.5004 0.4923 0.2429 0.4134 0.4134 25.1335 0.9143 0.9124 0.9132

Framework versions

  • Transformers 4.36.0
  • Pytorch 2.0.1+cu117
  • Datasets 2.14.5
  • Tokenizers 0.15.0
Downloads last month
3
Safetensors
Model size
570M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for GlycerinLOL/LLM_Teached_Pegasus_100k

Finetuned
(28)
this model