Modelpublic

snskrt/gemma-3-4b-it-sanskrit-ocr

sourceHugging Facegemmaupdated 1y agoView on Hugging Face

0likes25downloads

Model Card

gemma3-4B-sanskrit-ocr-lora

This model is a fine-tuned version of google/gemma-3-4b-it on the diabolic6045/sanskrit-ocr-parallel-corpus-chat-template dataset. This data is converted from snskrt/Sanskrit_OCR_Parallel_Corpus by Sanskrit Datasets. It achieves the following results on the evaluation set:

—Loss: 3.3256
—Memory/max Mem Active(gib): 11.52
—Memory/max Mem Allocated(gib): 11.52
—Memory/device Mem Reserved(gib): 12.26

Model description

#Todo

Training

Training Dataset

The model was trained on the diabolic6045/sanskrit-ocr-parallel-corpus-chat-template dataset, which contains Sanskrit text images paired with their corresponding transcriptions. The dataset was converted from the original snskrt/Sanskrit_OCR_Parallel_Corpus and formatted with chat templates for vision-language training.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

—learning_rate: 0.0002
—trainbatchsize: 4
—evalbatchsize: 4
—seed: 42
—distributed_type: multi-GPU
—num_devices: 2
—gradientaccumulationsteps: 4
—totaltrainbatch_size: 32
—totalevalbatch_size: 8
—optimizer: Use OptimizerNames.ADAMWBNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: cosine
—lrschedulerwarmup_steps: 55
—training_steps: 553

Training results

Training Loss	Epoch	Step	Validation Loss	Mem Active(gib)	Mem Allocated(gib)	Mem Reserved(gib)
No log	0	0	8.3469	9.34	9.34	9.42
3.5606	1.0	185	3.4639	11.52	11.52	12.26
2.839	2.0	370	3.3256	11.52	11.52	12.26

<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/> <details><summary>See axolotl config</summary>

axolotl version: 0.12.2

yaml

base_model: google/gemma-3-4b-it

# load_in_4bit: true  # Disabled for LoRA training

# gemma3 doesn't seem to play nice with ddp
ddp_find_unused_parameters: true

chat_template: gemma3
eot_tokens:
  - <end_of_turn>
datasets:
  - path: sanskrit_multimodal_train.json
    type: chat_template
    field_messages: messages


dataset_prepared_path: last_run_prepared
val_set_size: 0.01
output_dir: ./outputs/out-gemma3-4B

adapter: lora
lora_model_dir:

sequence_len: 2048
sample_packing: false


lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_modules: 'model.language_model.layers.[\d]+.(mlp|cross_attn|self_attn).(up|down|gate|q|k|v|o)_proj'

wandb_project: Sanskrit-OCR
wandb_entity:
wandb_watch:
wandb_name: gemma3-4B-sanskrit-ocr
wandb_log_model:
hub_model_id: diabolic6045/gemma3-4B-sanskrit-ocr-lora

gradient_accumulation_steps: 4
micro_batch_size: 4
num_epochs: 3
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

bf16: true
fp16:
tf32: true

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
logging_steps: 1
flash_attention: true
eager_attention: 

warmup_ratio: 0.1
evals_per_epoch: 1
saves_per_epoch: 1
weight_decay: 0.0

# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

</details><br>

Framework versions

—PEFT 0.17.0
—Transformers 4.55.2
—Pytorch 2.7.1+cu128
—Datasets 4.0.0
—Tokenizers 0.21.2