adapter: qlora
base_model: NousResearch/Yarn-Llama-2-7b-64k
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 74aeb5c63551d548_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/74aeb5c63551d548_train_data.json
  type:
    field_input: ''
    field_instruction: name
    field_output: text
    format: '{instruction}'
    no_input_format: '{instruction}'
    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: 1
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: error577/52788321-7bbb-4564-b1f3-e85bcf6b1f1d
hub_repo: null
hub_strategy: end
hub_token: null
learning_rate: 0.0001
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
max_steps: 300
micro_batch_size: 1
mlflow_experiment_name: /tmp/74aeb5c63551d548_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 12
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
saves_per_epoch: 1
sequence_len: 1024
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.02
wandb_entity: null
wandb_mode: online
wandb_name: 05d6a97d-a07b-424f-b289-aea33906250f
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 05d6a97d-a07b-424f-b289-aea33906250f
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

52788321-7bbb-4564-b1f3-e85bcf6b1f1d

This model is a fine-tuned version of NousResearch/Yarn-Llama-2-7b-64k on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.6145

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
gradient_accumulation_steps: 8
total_train_batch_size: 8
optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 10
training_steps: 300

Training results

Training Loss	Epoch	Step	Validation Loss
14.7383	0.0015	1	1.9144
8.0433	0.0367	25	0.9946
6.8449	0.0735	50	0.8665
6.4117	0.1102	75	0.8008
6.4906	0.1470	100	0.7581
5.5423	0.1837	125	0.7179
5.578	0.2205	150	0.6876
5.5843	0.2572	175	0.6619
6.4463	0.2940	200	0.6412
5.6181	0.3307	225	0.6261
5.3514	0.3674	250	0.6183
5.4112	0.4042	275	0.6149
5.4528	0.4409	300	0.6145

Framework versions

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