robinhad/UAlpaca-2.0-Mistral-7B

See axolotl config

axolotl version: 0.4.1

base_model: mistralai/Mistral-7B-v0.1
model_type: MistralForCausalLM
tokenizer_type: LlamaTokenizer

load_in_8bit: false
load_in_4bit: true
strict: false

chat_template: chatml

datasets:
  - path: /home/paniv/Projects/ualpaca2.json
    type: chat_template
    chat_template: chatml
    field_messages: conversations
    message_field_role: role
    message_field_content: content
    roles:
      user:
        - user
      assistant:
        - assistant


dataset_prepared_path: last_run_prepared
shuffle_merged_datasets: true
val_set_size: 0.02
output_dir: ./outputs/lora-out

adapter: qlora
lora_model_dir:

sequence_len: 8192
sample_packing: true
pad_to_sequence_len: true

lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj

wandb_project: UAlpaca2
wandb_entity:
wandb_watch:
wandb_name: full_train
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 5
num_epochs: 1
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: true
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3

warmup_steps: 10
evals_per_epoch: 4
eval_table_size:
eval_max_new_tokens: 128
eval_sample_packing: false
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:

outputs/lora-out

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5696

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: 0.0002
• train_batch_size: 5
• eval_batch_size: 5
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 4
• total_train_batch_size: 40
• total_eval_batch_size: 10
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.3714	0.0091	1	2.5733
1.1049	0.2551	28	0.6542
1.0633	0.5103	56	0.5824
1.0023	0.7654	84	0.5696

Framework versions

• PEFT 0.11.1
• Transformers 4.42.3
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1

Attribution

ELEKS supported this project through a grant dedicated to the memory of Oleksiy Skrypnyk.