ForeverBlue/EcoWitness-Qwen3.5-9B-RL-LoRA
image-text-to-text model by ForeverBlue
Older snapshot: check the original source for current information.
From the publisher
Model card & documentation
Source previewRead the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.
A LoRA adapter for Qwen/Qwen3.5-9B, trained on the EcoWitness witness protocol: multiple-choice questions about river ecology (Sentinel-2 image, site record, environmental-DNA species list) in which every piece of evidence carries an identifier and the model must answer with three lines, answer, support and loadbearing, naming the evidence it used. A rule engine recomputes the correct answer and the exact evidence sets for any evidence state, so every reply can be scored exactly. Code, method description and evaluation: (commit d262c98, ecowitness/rl.py, configs/rl.yaml). Each step, per GPU: 16 episodes (an original question, a silently edited evidence state, or an announced revision, mixed 0.4 / 0.35 / 0.25); 4 replies sampled per episode at temperature 1; each reply scored by the rule engine; advantage = reward minus the mean reward of that episode's samples; with probability 0.5 the reference reply is added as one more candidate with advantage 1 minus the mean sampled…
Read the full model card ↗ · Preview checked 2026-11-13T08:08:23.786Z
Inside the original model card — Document outline
- EcoWitness: Qwen3.5-9B LoRA trained with verifier-guided reinforcement learning
- Training
- Validation during training
- Use
- Limits
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Documentation belongs to its respective authors. Reported project/model license: apache-2.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
Model overview
ForeverBlue/EcoWitness-Qwen3.5-9B-RL-LoRA is a image-text-to-text model repository published by ForeverBlue on Hugging Face. The source reports the peft library.
This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.
Model facts
- Task
- image-text-to-text
- Library
- peft
- Recent downloads (30 days)
- 0
- Cumulative likes
- 0
- Hugging Face trending score
- 0
- Reported safetensors parameters
- Not reported
- Architecture
- Not reported
- License
- apache-2.0
- Access
- Not gated by Hugging Face
- Created
- 2026-11-13T06:39:33.000Z
- Last modified
- 2026-11-13T06:39:39.000Z
Compatibility and lineage
Reported languages: en
Reported base models: Qwen/Qwen3.5-9B
Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.
Use and evaluate
Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.
No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.
Source and freshness
Source: Hugging Face Hub. Metadata observed 2026-11-13T06:57:58.356Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.