See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: upstage/SOLAR-10.7B-Instruct-v1.0
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - f2e423591ee689ad_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/f2e423591ee689ad_train_data.json
  type:
    field_input: id
    field_instruction: constraints
    field_output: prompt
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
ddp_timeout: 1800
debug: null
deepspeed: null
device_map: auto
do_eval: true
early_stopping_patience: 3
eval_batch_size: 4
eval_max_new_tokens: 128
eval_steps: 150
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
gradient_checkpointing_kwargs:
  use_reentrant: true
group_by_length: true
hub_model_id: auxyus/56ca1e68-0244-41fa-80aa-e48ba641f2bf
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: 0
logging_steps: 10
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 16
lora_target_linear: true
lr_scheduler: constant
max_grad_norm: 1.0
max_memory:
  0: 75GB
max_steps: 1800
micro_batch_size: 4
mlflow_experiment_name: /tmp/f2e423591ee689ad_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 10
optim_args:
  adam_beta1: 0.9
  adam_beta2: 0.999
  adam_epsilon: 1e-08
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
relora_prune_ratio: 0.9
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 150
saves_per_epoch: null
sequence_len: 512
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: acopia-grant
wandb_mode: online
wandb_name: 846f4039-37f2-4763-b6ea-c8843ea17600
wandb_project: Gradients-On-191
wandb_run: your_name
wandb_runid: 846f4039-37f2-4763-b6ea-c8843ea17600
warmup_steps: 50
weight_decay: 0.0
xformers_attention: null

56ca1e68-0244-41fa-80aa-e48ba641f2bf

This model is a fine-tuned version of upstage/SOLAR-10.7B-Instruct-v1.0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2273

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.999,adam_epsilon=1e-08
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 50
• training_steps: 1800

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0009	1	2.2381
1.4787	0.1289	150	1.3686
1.3753	0.2579	300	1.3086
1.3978	0.3868	450	1.2788
1.3398	0.5158	600	1.2492
1.3674	0.6447	750	1.2322
1.3712	0.7737	900	1.2257
1.3378	0.9026	1050	1.2148
0.9355	1.0316	1200	1.2238
0.9736	1.1605	1350	1.2331
0.9775	1.2895	1500	1.2273

Framework versions

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