metadata
license: apache-2.0
library_name: peft
tags:
- generated_from_trainer
base_model: togethercomputer/evo-1-8k-base
model-index:
- name: lora_evo_ta_all_layers_10
results: []
lora_evo_ta_all_layers_10
This model is a fine-tuned version of togethercomputer/evo-1-8k-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.9544
Model description
Good model with A100
lora_alpha = 512 <---- 2x than that in model 9 (total 4x)
lora_dropout = 0.05
lora_r = 128 <---- same as model 9
epochs = 3
learning rate = 3e-4
warmup_steps=10
gradient_accumulation_steps = 8
train_batch = 1
eval_batch = 1
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: 0.0003
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_steps: 10
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 3.0259 | 1.0 | 266 | 2.9633 |
| 2.8429 | 2.0 | 532 | 2.9493 |
| 2.688 | 3.0 | 798 | 2.9544 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1