ViMath-PAL-deepseek-math-7B-LORA

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>

<details><summary>See axolotl config</summary>

axolotl version: 0.4.1

base_model: deepseek-ai/deepseek-math-7b-rl
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: false
strict: false
 
datasets:
  - path: /workspace/axolotl/vinh/PAL/input_output_dsmath.json
    type: input_output
dataset_prepared_path:
val_set_size: 0.05
eval_sample_packing: false
output_dir: /workspace/axolotl/vinh/deepseek-ai_deepseek-math-7b-rl-lora-2024-07-01-15-56-34

sequence_len: 2048
sample_packing: false
pad_to_sequence_len: false

adapter: lora
lora_model_dir: 
lora_r: 64
lora_alpha: 128
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 128
micro_batch_size: 1
num_epochs: 3
optimizer: paged_adamw_32bit
lr_scheduler: cosine
learning_rate: 2e-4

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention: 
flash_attention: true
s2_attention:

loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3

warmup_steps: 10
evals_per_epoch: 10
eval_table_size:
eval_max_new_tokens: 512
saves_per_epoch: 2
save_total_limit: 20
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
   pad_token: <｜end▁of▁sentence｜>

</details><br>

workspace/axolotl/vinh/deepseek-ai_deepseek-math-7b-rl-lora-2024-07-01-15-56-34

This model is a fine-tuned version of deepseek-ai/deepseek-math-7b-rl on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0276

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 128
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 3

Training results

Framework versions

• PEFT 0.11.1
• Transformers 4.41.1
• Pytorch 2.1.2+cu118
• Datasets 2.19.1
• Tokenizers 0.19.1