C3DS

CARDS-Qwen3.6-27B-API

README

License: apache-2.0

Results

Evaluated on the held-out CARDS test set (1,436 samples, Level 1, min_support ≥ 3):

Table with columns: Metric, Qwen3.5-27B, Qwen3.5-27B FT, Qwen3.6-27B FT, Claude Opus 4.6, Claude Opus 4.7
Metric	Qwen3.5-27B	Qwen3.5-27B FT	Qwen3.6-27B FT	Claude Opus 4.6	Claude Opus 4.7
Samples F1	0.844	0.884	0.893	0.893	0.882
Macro F1	0.710	0.766	0.748	0.751	0.771
Micro F1	0.854	0.877	0.885	0.881	0.874
Precision	0.870	0.879	0.893	0.863	0.868
Recall	0.838	0.874	0.876	0.900	0.880
Parse failures	86 / 1436	0 / 1436	2 / 1436	0 / 1436	0 / 1436

Ties Claude Opus 4.6 on Samples F1 at L1.
Wins Micro F1 and Precision at every hierarchy level.
Trails Claude Opus on Recall and Macro F1 for rare labels.
Parse failures drop from 6% (plain Qwen3.5-27B) to ~0 with fine-tuning.

Usage

With vLLM

bash

vllm serve C3DS/CARDS-Qwen3.6-27B \  --port 8000 \  --max-model-len 4096 \  --dtype bfloat16 \  --enable-prefix-caching \  --served-model-name CARDS-Qwen3.6-27B

Then query with any OpenAI-compatible client. The system prompt (slim_system_instruction) and the user-message suffix (cot_trigger) the model was trained with are bundled in this repo as cards_prompts.json — self-contained, with the CARDS taxonomy already inlined.

python

import jsonfrom huggingface_hub import hf_hub_downloadfrom openai import OpenAI prompts = json.load(open(hf_hub_download("C3DS/CARDS-Qwen3.6-27B", "cards_prompts.json")))slim_system_instruction = prompts["slim_system_instruction"]cot_trigger             = prompts["cot_trigger"] client = OpenAI(base_url="http://localhost:8000/v1", api_key="dummy") def classify(text):    resp = client.chat.completions.create(        model="CARDS-Qwen3.6-27B",        messages=[            {"role": "system", "content": slim_system_instruction},            {"role": "user",   "content": f"### Text:\n{text}\n\n{cot_trigger}"},        ],        temperature=0,        max_tokens=4000,    )    return resp.choices[0].message.content print(classify("These are only a few renewable energy technologies at work"))

The model produces a reasoning trace inside <think>…</think> followed by a YAML categories: block listing predicted CARDS codes. To parse: take the content after </think> and read the categories: list.

See the project repository for training scripts, evaluation code, and dataset preparation.

Multimodal — image + text

The base Qwen3.5/3.6 family supports image inputs via the OpenAI-compatible image_url content part, and this fine-tune preserves that capability — pass the system prompt below alongside an image (with or without caption text) and the model will classify the depicted claim under the CARDS taxonomy.

Serve vLLM with multimodal flags enabled:

bash

vllm serve C3DS/CARDS-Qwen3.6-27B \  --port 8000 \  --max-model-len 8192 \  --trust-remote-code \  --limit-mm-per-prompt image=4 \  --enable-prefix-caching \  --served-model-name CARDS-Qwen3.6-27B

python

import base64, json, mimetypesfrom pathlib import Pathfrom huggingface_hub import hf_hub_downloadfrom openai import OpenAI prompts = json.load(open(hf_hub_download("C3DS/CARDS-Qwen3.6-27B", "cards_prompts.json")))slim_system_instruction = prompts["slim_system_instruction"]cot_trigger             = prompts["cot_trigger"] def image_part(path):    p = Path(path)    mime = mimetypes.guess_type(p)[0] or "image/png"    b64 = base64.b64encode(p.read_bytes()).decode()    return {"type": "image_url", "image_url": {"url": f"data:{mime};base64,{b64}"}} client = OpenAI(base_url="http://localhost:8000/v1", api_key="dummy") resp = client.chat.completions.create(    model="CARDS-Qwen3.6-27B",    messages=[        {"role": "system", "content": slim_system_instruction},        {"role": "user", "content": [            {"type": "text", "text": "Read the image (and any caption below) and classify the climate claim it makes."},            image_part("screenshot.png"),            {"type": "text", "text": f"### Caption:\n<optional caption>\n\n{cot_trigger}"},        ]},    ],    temperature=0,    max_tokens=4000,)print(resp.choices[0].message.content)

Training

Base model: Qwen/Qwen3.6-27B
Method: LoRA (rank 16, α 16, dropout 0) on q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj, then merged into base weights
Dataset: C3DS/cards_sft_dataset
Framework: Unsloth + TRL SFTTrainer
Hyperparameters: 3 epochs, per_device_train_batch_size=1, gradient_accumulation_steps=8, lr=2e-4, cosine schedule, 10 warmup steps, max_seq_length=4096, adamw_8bit, bf16

Limitations

Macro F1 on rare labels. The model trails Claude Opus at L3 macro-F1 (0.483 vs 0.531), reflecting the long-tailed label distribution of CARDS.
Thinking tokens. Training used enable_thinking=True. Either parse output after </think>, or disable thinking at inference via chat_template_kwargs={"enable_thinking": false}. Reserve token budget for the reasoning trace before the final YAML block.

Citation

bibtex

@article{coan2026llm,  title     = {Large language model reveals an increase in climate contrarian speech in the United States Congress},  author    = {Coan, Travis G. and Malla, Ranadheer and Nanko, Mirjam O. and Kattrup, William and Roberts, J. Timmons and Cook, John and Boussalis, Constantine},  journal   = {Communications Sustainability},  volume    = {1},  number    = {1},  pages     = {37},  year      = {2026},  publisher = {Nature Publishing Group},  doi       = {10.1038/s44458-025-00029-z},  url       = {https://www.nature.com/articles/s44458-025-00029-z}}

License

Apache 2.0, inherited from Qwen3.6-27B.

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Model Details

Model Provider

C3DS

Model Tree

Base

Qwen/Qwen3.6-27B

Fine-tuned

this model

Input Modalities

TextImageVideo

Output Modalities

Text

Supported Functionality

Dedicated Endpoints

Explore FriendliAI today

Get started Talk to an engineer

README

License: apache-2.0

Results

Evaluated on the held-out CARDS test set (1,436 samples, Level 1, min_support ≥ 3):

Table with columns: Metric, Qwen3.5-27B, Qwen3.5-27B FT, Qwen3.6-27B FT, Claude Opus 4.6, Claude Opus 4.7
Metric	Qwen3.5-27B	Qwen3.5-27B FT	Qwen3.6-27B FT	Claude Opus 4.6	Claude Opus 4.7
Samples F1	0.844	0.884	0.893	0.893	0.882
Macro F1	0.710	0.766	0.748	0.751	0.771
Micro F1	0.854	0.877	0.885	0.881	0.874
Precision	0.870	0.879	0.893	0.863	0.868
Recall	0.838	0.874	0.876	0.900	0.880
Parse failures	86 / 1436	0 / 1436	2 / 1436	0 / 1436	0 / 1436

Ties Claude Opus 4.6 on Samples F1 at L1.
Wins Micro F1 and Precision at every hierarchy level.
Trails Claude Opus on Recall and Macro F1 for rare labels.
Parse failures drop from 6% (plain Qwen3.5-27B) to ~0 with fine-tuning.

Usage

With vLLM

bash

vllm serve C3DS/CARDS-Qwen3.6-27B \  --port 8000 \  --max-model-len 4096 \  --dtype bfloat16 \  --enable-prefix-caching \  --served-model-name CARDS-Qwen3.6-27B

python

import jsonfrom huggingface_hub import hf_hub_downloadfrom openai import OpenAI prompts = json.load(open(hf_hub_download("C3DS/CARDS-Qwen3.6-27B", "cards_prompts.json")))slim_system_instruction = prompts["slim_system_instruction"]cot_trigger             = prompts["cot_trigger"] client = OpenAI(base_url="http://localhost:8000/v1", api_key="dummy") def classify(text):    resp = client.chat.completions.create(        model="CARDS-Qwen3.6-27B",        messages=[            {"role": "system", "content": slim_system_instruction},            {"role": "user",   "content": f"### Text:\n{text}\n\n{cot_trigger}"},        ],        temperature=0,        max_tokens=4000,    )    return resp.choices[0].message.content print(classify("These are only a few renewable energy technologies at work"))

See the project repository for training scripts, evaluation code, and dataset preparation.

Multimodal — image + text

Serve vLLM with multimodal flags enabled:

bash

vllm serve C3DS/CARDS-Qwen3.6-27B \  --port 8000 \  --max-model-len 8192 \  --trust-remote-code \  --limit-mm-per-prompt image=4 \  --enable-prefix-caching \  --served-model-name CARDS-Qwen3.6-27B

python

import base64, json, mimetypesfrom pathlib import Pathfrom huggingface_hub import hf_hub_downloadfrom openai import OpenAI prompts = json.load(open(hf_hub_download("C3DS/CARDS-Qwen3.6-27B", "cards_prompts.json")))slim_system_instruction = prompts["slim_system_instruction"]cot_trigger             = prompts["cot_trigger"] def image_part(path):    p = Path(path)    mime = mimetypes.guess_type(p)[0] or "image/png"    b64 = base64.b64encode(p.read_bytes()).decode()    return {"type": "image_url", "image_url": {"url": f"data:{mime};base64,{b64}"}} client = OpenAI(base_url="http://localhost:8000/v1", api_key="dummy") resp = client.chat.completions.create(    model="CARDS-Qwen3.6-27B",    messages=[        {"role": "system", "content": slim_system_instruction},        {"role": "user", "content": [            {"type": "text", "text": "Read the image (and any caption below) and classify the climate claim it makes."},            image_part("screenshot.png"),            {"type": "text", "text": f"### Caption:\n<optional caption>\n\n{cot_trigger}"},        ]},    ],    temperature=0,    max_tokens=4000,)print(resp.choices[0].message.content)

Training

Base model: Qwen/Qwen3.6-27B
Method: LoRA (rank 16, α 16, dropout 0) on q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj, then merged into base weights
Dataset: C3DS/cards_sft_dataset
Framework: Unsloth + TRL SFTTrainer
Hyperparameters: 3 epochs, per_device_train_batch_size=1, gradient_accumulation_steps=8, lr=2e-4, cosine schedule, 10 warmup steps, max_seq_length=4096, adamw_8bit, bf16

Limitations

Macro F1 on rare labels. The model trails Claude Opus at L3 macro-F1 (0.483 vs 0.531), reflecting the long-tailed label distribution of CARDS.
Thinking tokens. Training used enable_thinking=True. Either parse output after </think>, or disable thinking at inference via chat_template_kwargs={"enable_thinking": false}. Reserve token budget for the reasoning trace before the final YAML block.

Citation

bibtex

@article{coan2026llm,  title     = {Large language model reveals an increase in climate contrarian speech in the United States Congress},  author    = {Coan, Travis G. and Malla, Ranadheer and Nanko, Mirjam O. and Kattrup, William and Roberts, J. Timmons and Cook, John and Boussalis, Constantine},  journal   = {Communications Sustainability},  volume    = {1},  number    = {1},  pages     = {37},  year      = {2026},  publisher = {Nature Publishing Group},  doi       = {10.1038/s44458-025-00029-z},  url       = {https://www.nature.com/articles/s44458-025-00029-z}}

License

Apache 2.0, inherited from Qwen3.6-27B.