---
library_name: transformers
pipeline_tag: text-generation
base_model: Qwen/Qwen3-8B-Base
datasets:
- CompassioninMachineLearning/urban_12738_cleaned
tags:
- continued-pretraining
- bf16
---

# Qwen3-8b-urban-qwen-seed1-20260930-CPT-merged-epoch-3

Standalone merged BF16 model from epoch 3.0, step 1134.

Dataset: `CompassioninMachineLearning/urban_12738_cleaned` at `ef7c0e742df63ea319e35d02d9f6ba63d6e7c68d`.

Training: 10,072 distinct documents plus 2,000 repeat exposures per epoch; 200 disjoint validation documents.

Merged with Unsloth's native `save_pretrained_merged(save_method="merged_16bit")`. Weights are validated BF16 and packaged losslessly into eight safetensors shards. No adapter is required to load this model.

See `run_manifest.json` for base revision, document selection hashes, training parameters and export validation. Training does not establish an improvement in compassion; evaluate that separately.