base_model: google/gemma-3-1b-it
# optionally might have model_type or tokenizer_type
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
# Automatically upload checkpoint and final model to HF
# hub_model_id: username/custom_model_name

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

load_in_8bit: false
load_in_4bit: false

# huggingface repo
chat_template: gemma3
eot_tokens:
  - <end_of_turn>
datasets:
  - path: sam2ai/en-or-hi-ml-bn-translation
    type: chat_template
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
    roles:
      assistant:
        - gpt
      user:
        - human

val_set_size: 0.1
output_dir: ./outputs/gemma-3-1b-wat2025

  #adapter: qlora
  #lora_r: 32
  #lora_alpha: 16
  #lora_dropout: 0.05
  #lora_target_linear: true

sequence_len: 4096
sample_packing: true
eval_sample_packing: false


wandb_project: gemma3-en-indic-wat2025
wandb_entity:
wandb_watch:
wandb_name: gemma3-1b
wandb_log_model:


gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 4
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

bf16: auto
tf32: false

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

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

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

outputs/gemma-3-1b-wat2025

This model is a fine-tuned version of google/gemma-3-1b-it on the sam2ai/en-or-hi-ml-bn-translation dataset.

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: 2
eval_batch_size: 2
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 4
total_train_batch_size: 64
total_eval_batch_size: 16
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: 82
training_steps: 828

Training results

Framework versions

Transformers 4.55.2
Pytorch 2.7.0+gitf717b2a
Datasets 4.0.0
Tokenizers 0.21.4

sam2ai/gemma3-1b-en-od-ml-hi-bn

outputs/gemma-3-1b-wat2025

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

sam2ai/
gemma3-1b-en-od-ml-hi-bn