Theoistic/gemma-3-1b-fo

Hugging Face
TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:1BQuant:BF16Context Size:32kPublished:Nov 22, 2025License:otherArchitecture:Transformer Featherless Exclusive Warm

Theoistic/gemma-3-1b-fo is a 1 billion parameter Gemma-based causal language model, fine-tuned from google/gemma-3-1b-pt. This model is specifically optimized for translation tasks, leveraging its fine-tuning on a dedicated translation dataset. It is designed for applications requiring efficient and accurate language translation.

Loading preview...

Model Overview

Theoistic/gemma-3-1b-fo is a 1 billion parameter language model, fine-tuned from the google/gemma-3-1b-pt base model. This iteration has undergone specific fine-tuning on a translation dataset, indicating its primary specialization in language translation tasks.

Key Characteristics

  • Base Model: Fine-tuned from Google's Gemma-3-1B-PT architecture.
  • Parameter Count: 1 billion parameters, offering a balance between performance and computational efficiency.
  • Context Length: Supports a context length of 32768 tokens.
  • Specialization: Optimized for translation, based on its training on a dedicated translation dataset.

Training Details

The model was trained with the following hyperparameters:

  • Learning Rate: 5e-05
  • Batch Size: A train_batch_size of 16 and eval_batch_size of 8, with a gradient_accumulation_steps of 4, resulting in a total_train_batch_size of 64.
  • Optimizer: ADAMW_TORCH with default betas and epsilon.
  • Scheduler: Cosine learning rate scheduler.
  • Epochs: Trained for 2.0 epochs.

Intended Use Cases

This model is particularly suitable for applications requiring efficient and accurate translation capabilities, given its specialized fine-tuning. Developers can leverage it for integrating translation features into various platforms and services.

Finetunes
1 models