See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: princeton-nlp/Sheared-LLaMA-1.3B
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - cffeb2068f28cc6e_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/cffeb2068f28cc6e_train_data.json
  type:
    field_input: main_domain
    field_instruction: title
    field_output: description
    format: '{instruction} {input}'
    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: 2
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 32
gradient_checkpointing: false
max_grad_norm: 1
group_by_length: false
hub_model_id: error577/f5007068-224b-49a6-8aff-b438e87bf3ef
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 10
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_steps: 500
micro_batch_size: 1
mlflow_experiment_name: /tmp/cffeb2068f28cc6e_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 8
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: 512
special_tokens:
  pad_token: </s>
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: d6d65c6a-7a9e-4819-9e86-a9ea4b9f10c7
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: d6d65c6a-7a9e-4819-9e86-a9ea4b9f10c7
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

f5007068-224b-49a6-8aff-b438e87bf3ef

This model is a fine-tuned version of princeton-nlp/Sheared-LLaMA-1.3B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.0399

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: 32
• total_train_batch_size: 32
• 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: 500

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0068	1	2.4177
2.2864	0.2173	32	2.1603
2.0811	0.4345	64	2.1125
2.1192	0.6518	96	2.0913
2.1414	0.8691	128	2.0728
1.9297	1.0864	160	2.0627
1.9738	1.3036	192	2.0563
1.907	1.5209	224	2.0506
2.0121	1.7382	256	2.0443
1.8795	1.9554	288	2.0390
1.9241	2.1727	320	2.0422
1.776	2.3900	352	2.0423
1.8113	2.6073	384	2.0412
1.7836	2.8245	416	2.0393
1.8139	3.0418	448	2.0393
1.7138	3.2591	480	2.0399

Framework versions

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