talzoomanzoo/uid_gated_aime_qwen3-1-7b_ep1
The talzoomanzoo/uid_gated_aime_qwen3-1-7b_ep1 is a 1.7 billion parameter language model based on the Qwen3 architecture, developed by talzoomanzoo. This model is a merged full-weight checkpoint of Qwen/Qwen3-1.7B, enhanced with a LoRA adapter from UID-gated GRPO AIME training. It features a 32768 token context length and is specifically adapted for tasks benefiting from its unique training methodology.
Loading preview...
Model Overview
The talzoomanzoo/uid_gated_aime_qwen3-1-7b_ep1 is a 1.7 billion parameter language model built upon the Qwen/Qwen3-1.7B base architecture. This specific iteration represents a merged full-weight checkpoint, integrating a LoRA (Low-Rank Adaptation) adapter. The adapter, with a rank of 64 and alpha of 32, originates from UID-gated GRPO AIME training, specifically from epoch 1 at global_step_8.
Key Characteristics
- Base Model: Qwen/Qwen3-1.7B, providing a robust foundation.
- Parameter Count: 1.7 billion parameters, offering a balance of performance and efficiency.
- Context Length: Supports a substantial context window of 32768 tokens.
- Training Method: Enhanced through a LoRA adapter derived from UID-gated GRPO AIME training, indicating a specialized fine-tuning approach.
Potential Use Cases
This model is suitable for applications that can leverage its Qwen3 base and the specific adaptations from the UID-gated GRPO AIME training. Developers looking for a 1.7B parameter model with a large context window and specialized fine-tuning might find this model particularly useful for tasks aligned with its training objectives.