See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: NousResearch/Yarn-Mistral-7b-64k
bf16: auto
chat_template: llama3
cosine_min_lr_ratio: 0.1
data_processes: 16
dataset_prepared_path: null
datasets:
- data_files:
  - f5b7dd3a5eab14ca_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/f5b7dd3a5eab14ca_train_data.json
  type:
    field_input: original_abstract
    field_instruction: original_title
    field_output: processed_title
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
device_map: '{'''':torch.cuda.current_device()}'
do_eval: true
early_stopping_patience: 1
eval_batch_size: 1
eval_sample_packing: false
eval_steps: 25
evaluation_strategy: steps
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 64
gradient_checkpointing: true
group_by_length: true
hub_model_id: sn56m2/cd18b217-8ca2-4ef8-9e7a-f35784db2082
hub_repo: stevemonite
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lora_target_modules:
- q_proj
- k_proj
- v_proj
- o_proj
- gate_proj
- down_proj
- up_proj
lr_scheduler: cosine
max_grad_norm: 1.0
max_memory:
  0: 70GiB
max_steps: 457
micro_batch_size: 1
mlflow_experiment_name: /tmp/f5b7dd3a5eab14ca_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optim_args:
  adam_beta1: 0.9
  adam_beta2: 0.95
  adam_epsilon: 1e-5
optimizer: adamw_torch
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 50
save_strategy: steps
sequence_len: 2056
special_tokens:
  pad_token: </s>
strict: false
tf32: false
tokenizer_type: AutoTokenizer
torch_compile: false
train_on_inputs: false
trust_remote_code: true
val_set_size: 50
wandb_entity: sn56-miner
wandb_mode: disabled
wandb_name: cd18b217-8ca2-4ef8-9e7a-f35784db2082
wandb_project: god
wandb_run: nsef
wandb_runid: cd18b217-8ca2-4ef8-9e7a-f35784db2082
warmup_raio: 0.03
warmup_ratio: 0.04
weight_decay: 0.01
xformers_attention: null

cd18b217-8ca2-4ef8-9e7a-f35784db2082

This model is a fine-tuned version of NousResearch/Yarn-Mistral-7b-64k on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0005

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: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 64
• total_train_batch_size: 256
• total_eval_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.95,adam_epsilon=1e-5
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 18
• training_steps: 457

Training results

Training Loss	Epoch	Step	Validation Loss
64.3546	0.0026	1	1.3639
0.608	0.0642	25	0.0062
0.2007	0.1283	50	0.0058
0.0584	0.1925	75	0.0009
0.3828	0.2566	100	0.0014
0.2763	0.3208	125	0.0009
0.1823	0.3850	150	0.0012
0.1509	0.4491	175	0.0010
0.0494	0.5133	200	0.0003
0.0316	0.5774	225	0.0005

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1