See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: NousResearch/Yarn-Llama-2-7b-128k
bf16: auto
dataset_prepared_path: null
datasets:
- data_files:
  - cff7ac798e6d5dcd_train_data.json
  ds_type: json
  format: custom
  path: cff7ac798e6d5dcd_train_data.json
  preprocessing:
  - shuffle: true
  type:
    field: null
    field_input: input
    field_instruction: instruction
    field_output: response
    field_system: null
    format: null
    no_input_format: null
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 5
flash_attention: null
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: false
hub_model_id: taopanda/test-Yarn-Llama-2-7b-128k
learning_rate: 0.0004
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 128
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 150
micro_batch_size: 4
model_type: LlamaForCausalLM
num_epochs: 1
optimizer: adamw_torch
output_dir: ./outputs/lora-out/taopanda_test-Yarn-Llama-2-7b-128k
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
save_steps: 0.25
save_total_limit: 1
seed: 42
sequence_len: 1024
special_tokens: null
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.1
wandb_entity: fatcat87-taopanda
wandb_log_model: null
wandb_mode: online
wandb_name: taopanda_test-Yarn-Llama-2-7b-128k
wandb_project: subnet56-test
wandb_runid: taopanda_test-Yarn-Llama-2-7b-128k
wandb_watch: null
warmup_ratio: 0.06
weight_decay: 0.0
xformers_attention: null

test-Yarn-Llama-2-7b-128k

This model is a fine-tuned version of NousResearch/Yarn-Llama-2-7b-128k on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0518

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: 0.0004
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• total_eval_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 4
• training_steps: 81

Training results

Training Loss	Epoch	Step	Validation Loss
2.8167	0.0124	1	2.8724
1.6497	0.2112	17	1.5763
1.3523	0.4224	34	1.3288
0.9764	0.6335	51	1.1532
0.9666	0.8447	68	1.0518

Framework versions

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