21 lines
566 B
Markdown
21 lines
566 B
Markdown
---
|
|
license: apache-2.0
|
|
base_model: hamishivi/sft_qwen3_8b_our_sft
|
|
library_name: transformers
|
|
tags:
|
|
- open-instruct
|
|
- grpo
|
|
- qwen3
|
|
- swerl
|
|
---
|
|
|
|
# SWERL Qwen3 8B Endless Terminals GRPO
|
|
|
|
Step 500 checkpoint from the `hamishivi/agent-task-endless-terminals` GRPO run.
|
|
|
|
- Base model: `hamishivi/sft_qwen3_8b_our_sft`
|
|
- Run ID: `swerl_qwen3_8b_our_sft_agent_task_endless_terminals_grpo__seed42_direct_pytest_close_release_pool512_20260518_085210`
|
|
- Training completed: 2026-05-20
|
|
|
|
This checkpoint is intended for internal evaluation and continuation experiments.
|