You can deploy this model and start earning money today!

CheapsetZero/
89dc2c4b-66a4-4789-9cc0-8984667f3fae

Name: 89dc2c4b-66a4-4789-9cc0-8984667f3fae
Availability: InStock

generated_from_trainer

NousResearch/Genstruct-7B

adapter

apache-2.0

Views

No views yet

See axolotl config

axolotl version: 0.4.1

yaml

adapter: lora
base_model: NousResearch/Genstruct-7B
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - ec5c455c4564f1e8_train_data.json
  ds_type: json
  field: prompt
  path: /workspace/input_data/
  split: train
  type: completion
ddp_find_unused_parameters: false
debug: null
deepspeed: null
early_stopping_patience: null
ema_decay: 0.995
ema_update_after_step: 200
eps: 1.0e-06
eval_max_new_tokens: 256
eval_table_size: null
evals_per_epoch: 4
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 1
gradient_checkpointing: true
gradient_clipping: 0.5
gradient_normalization: true
greater_is_better: false
group_by_length: false
hub_model_id: CheapsetZero/89dc2c4b-66a4-4789-9cc0-8984667f3fae
learning_rate: 0.0002
load_best_model_at_end: true
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_nan_inf_filter: true
logging_steps: 1
lora_alpha: 128
lora_dropout: 0.1
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: 9350
metric_for_best_model: eval_loss
micro_batch_size: 24
min_lr: 4.0e-05
mlflow_experiment_name: /tmp/ec5c455c4564f1e8_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
reward_model_sampling_temperature: 0.7
s2_attention: null
sample_packing: false
save_total_limit: 3
saves_per_epoch: 4
sequence_len: 1024
skip_nan_gradients: true
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trl:
  adaptive_beta: true
  beta: 0.12
  entropy_coeff: 0.01
  gradient_normalization: true
  kl_monitoring: true
  max_completion_length: 1024
  num_generations: 12
  reward_funcs:
  - rewards_c805d5cb-b707-4539-935c-44eca16d0868.reward_short_sentences
  - rewards_c805d5cb-b707-4539-935c-44eca16d0868.reward_low_unique_words_percentage
  - rewards_c805d5cb-b707-4539-935c-44eca16d0868.reward_flesch_kincaid_grade
  - rewards_c805d5cb-b707-4539-935c-44eca16d0868.reward_high_difficult_words_percentage
  - rewards_c805d5cb-b707-4539-935c-44eca16d0868.reward_specific_char_count
  reward_weights:
  - 4.045081822319539
  - 3.884099775347396
  - 0.45526596240455186
  - 0.7509416657917278
  - 2.5786937124531386
  target_kl: 0.01
  use_vllm: false
trust_remote_code: true
use_ema: true
use_peft: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: c805d5cb-b707-4539-935c-44eca16d0868
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: c805d5cb-b707-4539-935c-44eca16d0868
warmup_steps: 428
weight_decay: 0.01
xformers_attention: null

89dc2c4b-66a4-4789-9cc0-8984667f3fae

This model is a fine-tuned version of NousResearch/Genstruct-7B on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.5272

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: 0.0002
train_batch_size: 24
eval_batch_size: 24
seed: 42
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: 428
training_steps: 2437

Training results

Training Loss	Epoch	Step	Validation Loss
1.8804	0.0012	1	2.0782
0.5421	0.2509	204	0.4991
0.7064	0.5018	408	0.5257
0.5176	0.7528	612	0.5374
0.3154	1.0037	816	0.5444
0.4984	1.2546	1020	0.5417
0.388	1.5055	1224	0.5338
0.3166	1.7565	1428	0.5176
0.1708	2.0074	1632	0.5176
0.2574	2.2583	1836	0.5314
0.1537	2.5092	2040	0.5294
0.2695	2.7601	2244	0.5272

Framework versions

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