adapter: lora
base_model: NousResearch/Yarn-Solar-10b-32k
bf16: true
chat_template: llama3
data_processes: 40
dataset_prepared_path: null
datasets:
- data_files:
  - 944428010357d088_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/944428010357d088_train_data.json
  type:
    field_input: description
    field_instruction: input persona
    field_output: synthesized text
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
device_map: auto
do_eval: true
early_stopping_patience: 4
eval_batch_size: 4
eval_max_new_tokens: 128
eval_steps: 250
eval_table_size: null
evals_per_epoch: null
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: true
hub_model_id: cilorku/3d6605d9-3015-4747-adc1-50c3cbccb457
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 9.0e-05
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 128
lora_dropout: 0.04
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lr_scheduler: cosine
lr_scheduler_warmup_steps: 50
max_grad_norm: 1.0
max_memory:
  0: 135GB
max_steps: 2000
micro_batch_size: 16
mlflow_experiment_name: /tmp/944428010357d088_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optim_args:
  adam_beta1: 0.9
  adam_beta2: 0.95
  adam_epsilon: 1e-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
save_steps: 500
saves_per_epoch: null
seed: 17333
sequence_len: 1024
special_tokens:
  pad_token: </s>
strict: false
tf32: true
tokenizer_type: AutoTokenizer
total_train_batch_size: 64
train_batch_size: 8
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 2c1e703b-be68-4ed7-980c-0f149857291b
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 2c1e703b-be68-4ed7-980c-0f149857291b
warmup_steps: 300
weight_decay: 0.0
xformers_attention: null

3d6605d9-3015-4747-adc1-50c3cbccb457

This model is a fine-tuned version of NousResearch/Yarn-Solar-10b-32k on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.7249

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: 9e-05
train_batch_size: 16
eval_batch_size: 4
seed: 17333
gradient_accumulation_steps: 4
total_train_batch_size: 64
optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.95,adam_epsilon=1e-8
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 300
training_steps: 2000

Training results

Training Loss	Epoch	Step	Validation Loss
2.9674	0.0014	1	1.1262
3.4306	0.3382	250	0.7594
3.567	0.6764	500	0.7319
2.0008	1.0145	750	0.6969
1.8881	1.3527	1000	0.6876
1.8083	1.6909	1250	0.6635
1.136	2.0291	1500	0.7033
1.0433	2.3673	1750	0.7137
1.1869	2.7054	2000	0.7249

Framework versions

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