GGUF builds of C3DS/CARDS-Qwen3.5-4B — Qwen3.5-4B fine-tuned for climate-contrarian-claim classification under the CARDS taxonomy from Coan et al. (2025).
For accuracy comparisons against the BF16 sibling, larger Qwen3 variants, Claude Opus 4.6 / 4.7 and GPT-5.5, see the BF16 model card. Headline number on the held-out CARDS test set: L1 samples F1 = 0.838 (1 / 1,436 parse failures).
Available files
File
Quantization
Approx. size
Use
Qwen3.5-4B.Q4_K_M.gguf
Q4_K_M (4-bit, mixed)
~2.8 GB
smallest; runs on consumer CPUs / low-VRAM GPUs
Qwen3.5-4B.Q8_0.gguf
Q8_0 (8-bit)
~4.8 GB
near-BF16 quality, slightly larger
Qwen3.5-4B.BF16-mmproj.gguf
BF16 mmproj
—
multimodal projector (vision encoder); pair with either weight file for image input
The Q4_K_M and Q8_0 files are alternatives — pick one based on your accuracy / memory trade-off. The mmproj file is additional and only needed for image inputs.
Usage
With Ollama
A pre-published Ollama version is available at exec3ds/cards-qwen3.5. Pull and run the 4B (Q4_K_M, ~2.7 GB):
ollama run exec3ds/cards-qwen3.5:4b
The same namespace also publishes :9b and :27b tags for the larger siblings. The bundled chat template is applied automatically — Qwen3's <think> block precedes the YAML categories: output.
Image input works too — pass an image path on the command line:
bash
1ollama run exec3ds/cards-qwen3.5:4b ./image.webp \2"What climate claim is being made in this image?"
Text-only with llama.cpp
bash
1llama-cli \2 -hf C3DS/CARDS-Qwen3.5-4B-GGUF:Q4_K_M \3 --jinja \4 -p "Classify the following climate claim under the CARDS taxonomy:\nThese are only a few renewable energy technologies at work."
The --jinja flag tells llama-cli to use the bundled chat template, which adds Qwen3's <think> reasoning trace plus the YAML categories: block at the end. Parse the content after </think> to extract the predicted CARDS codes.
For an OpenAI-compatible local server, use llama-server:
The base Qwen3.5 architecture (Qwen3_5ForConditionalGeneration) is multimodal. Pair the weight file with the BF16-mmproj.gguf projector and use llama-mtmd-cli:
bash
1llama-mtmd-cli \2 -hf C3DS/CARDS-Qwen3.5-4B-GGUF:Q4_K_M \3 --mmproj-hf C3DS/CARDS-Qwen3.5-4B-GGUF/Qwen3.5-4B.BF16-mmproj.gguf \4 --jinja \5 --image screenshot.png \6 -p "Read the image (and any caption) and classify the climate claim under the CARDS taxonomy."
For the canonical CARDS system prompt and CoT trigger, fetch
cards_prompts.json
from the BF16 sibling repo and pass slim_system_instruction as the system
prompt + cot_trigger at the end of the user turn.
Multimodal projector: BF16 (kept full-precision for accuracy on vision inputs)
Limitations
Thinking tokens. Training used enable_thinking=True. Either parse output after </think>, or disable thinking in your sampler / system prompt. Reserve token budget for the reasoning trace before the final YAML block.
Quantization trade-offs. Q4_K_M is the most aggressive; expect a modest drop from BF16 in the long-tailed L3 macro-F1 metrics. Q8_0 tracks BF16 more closely.
Multimodal use is opportunistic. Training was text-only; image-input behavior comes from the preserved base vision encoder. Performance on image inputs has not been quantitatively benchmarked.
Citation
bibtex
1@article{coan2025cards,
2 title = {Large language model reveals an increase in climate contrarian speech in the United States Congress},
3 author = {Coan, Travis G. and Malla, Ranadheer and Nanko, Mirjam O. and Kattrup, William and Roberts, J. Timmons and Cook, John and Boussalis, Constantine},
4 journal = {Communications Sustainability},
5 volume = {1},
6 pages = {37},
7 year = {2025},
8 doi = {10.1038/s44458-025-00029-z}
9}
