lora_evo_ta_all_layers_6

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: 3.1104

Model description

lora_alpha = 32

lora_dropout = 0.05

lora_r = 16

epochs = 3

learning rate = 3e-4

warmup_steps=0.5

gradient_accumulation_steps = 8

train_batch = 1

eval_batch = 1

Training of only last 40 linear modules [120:160] instead of [0:160] <------

This changes the # of trainable params to 8,914,944

Intended uses & limitations

More information needed

Training and evaluation data

look at files

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
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 0.5
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
3.1331	0.9925	33	3.1104
3.1331	1.9850	66	3.1104
3.0635	2.9774	99	3.1104

Framework versions

• PEFT 0.11.1
• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1