Downloads · 30 days
61
4% of all-time downloads
Ling-3.0-flash-REAP320-81B-A5B
bloomer010/Ling-3.0-flash-REAP320-81B-A5B
Ling-3.0-flash-REAP320-81B-A5B is a text generation model from bloomer010. Use it when you need the model to write or continue text. It is set up for transformers.
All-time downloads
1.5K
Public
Parameters
81B
162 GB on disk
Likes
1
Public
Repo makeup
Click a slice to open those files.
.safetensors162 GB · 100%
Parameter types
How the weights are stored.
BF1681B · 100%
Try a prompt
Base models
Model card
From the Hugging Face model README
This is an experimental REAP.
Ling-3.0-flash REAP320 (81B total / 5.1B active)
[320 of 512 routed experts kept per layer - 38% of experts pruned] from inclusionAI/Ling-3.0-flash (124B total / 5.1B active).
Method: one-shot REAP (Router-weighted Expert Activation Pruning) - experts scored by router-gate-value × output-L2-norm over calibration data, lowest-scoring deleted. No fine-tuning, no recovery training.
Calibration: 1M tokens, 50/25/25 ultrachat / wikitext / code
BF16 safetensors. Loads with trust_remote_code=True (custom bailing_hybrid / BailingMoeV3 code).
Research artifact - quantized builds live in the sibling -GGUF repo.
2026-08-21: chat_template.jinja updated to support reasoning_effort
(low = no thinking, high = thinking, default unchanged).
- Task
- Text Generation
- Library
- transformers
- Type
- bailing_hybrid
- Created
- Aug 15, 2026
- Updated
- Aug 21, 2026