FlameF0X/TerraLM-350M

TEXT GENERATIONConcurrency Cost:1Model Size:0.35BQuant:BF16Ctx Length:32kPublished:May 8, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

FlameF0X/TerraLM-350M is a 350 million parameter language model developed by FlameF0X, finetuned from LiquidAI/LFM2.5-350M. This model was trained using Unsloth and Huggingface's TRL library, enabling 2x faster finetuning. With a context length of 32768 tokens, it is optimized for efficient processing and deployment.

Loading preview...

TerraLM-350M: An Efficiently Finetuned Language Model

TerraLM-350M is a 350 million parameter language model developed by FlameF0X, building upon the LiquidAI/LFM2.5-350M base model. A key differentiator of this model is its training methodology, leveraging Unsloth and Huggingface's TRL library, which facilitated a 2x speedup in the finetuning process.

Key Characteristics

  • Parameter Count: 350 million parameters, offering a balance between performance and computational efficiency.
  • Base Model: Finetuned from LiquidAI/LFM2.5-350M.
  • Training Efficiency: Utilizes Unsloth and Huggingface TRL for accelerated finetuning.
  • Context Length: Supports a substantial context window of 32768 tokens.

Potential Use Cases

TerraLM-350M is suitable for applications requiring a compact yet capable language model, especially where rapid deployment and efficient resource utilization are critical. Its efficient finetuning process suggests it could be a good candidate for custom applications where domain-specific adaptation is needed without extensive training times.