Issue Station
Sign In
Tools hereToolbox

transformers - ✅(Solved) Fix Add AutoModelForSequenceClassification support for Qwen3.5 (Qwen3_5Config) [1 pull requests, 1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
huggingface/transformers#44405•Fetched 2026-04-08 00:28:41
View on GitHub
Comments
0
Participants
1
Timeline
4
Reactions
1
Author
medhakimbedhiefmedhakimbedhief
Participants
medhakimbedhiefmedhakimbedhief
Timeline (top)
closed ×1cross-referenced ×1labeled ×1subscribed ×1

Error Message

ValueError: Unrecognized configuration class <class 'transformers.models.qwen3_5.configuration_qwen3_5.Qwen3_5Config'> for this kind of AutoModel: AutoModelForSequenceClassification.

Root Cause

Transformers raises:

ValueError: Unrecognized configuration class <class 'transformers.models.qwen3_5.configuration_qwen3_5.Qwen3_5Config'>
for this kind of AutoModel: AutoModelForSequenceClassification.

This error occurs because Qwen3_5Config is not registered in the internal mapping of configuration classes to sequence-classification model classes. Sequence classification mappings currently include Qwen3Config and related, but not Qwen3_5Config. This prevents classification from loading via the auto class.

Fix Action

Fix / Workaround

Expected Behavior

AutoModelForSequenceClassification.from_pretrained("Qwen/Qwen3.5-0.8B", ...) should:

  • Recognize Qwen3_5Config
  • Instantiate a Qwen3_5ForSequenceClassification class (or equivalent)
  • Allow fine-tuning and inference on classification tasks without manual patching

PR fix notes

PR #44406: Add Qwen3.5 support for sequence classification

Description (problem / solution / changelog)

Adds sequence-classification support for Qwen3.5 in AutoModelForSequenceClassification.

What does this PR do? This PR enables loading Qwen3.5 checkpoints with AutoModelForSequenceClassification, which previously failed with: ValueError: Unrecognized configuration class Qwen3_5Config for AutoModelForSequenceClassification.

Changes

Why both mappings? Qwen3.5 uses a composite VLM config (qwen3_5) with a text sub-config (qwen3_5_text). Registering both ensures classification works for direct text config usage and composite config loading paths.

Before : Loading from e.g. Qwen/Qwen3.5-0.8B raises ValueError: Unrecognized configuration class Qwen3_5Config for AutoModelForSequenceClassification.

After this PR: Loading from Qwen/Qwen3.5-0.8B now resolves to Qwen3_5ForSequenceClassification.

@Cyrilvallez @zucchini-nlp @ArthurZucker

Fixes #44405

Changed files

  • docs/source/en/model_doc/qwen3_5.md (modified, +5/-0)
  • src/transformers/models/auto/modeling_auto.py (modified, +2/-0)
  • src/transformers/models/qwen3_5/modeling_qwen3_5.py (modified, +6/-1)
  • src/transformers/models/qwen3_5/modular_qwen3_5.py (modified, +6/-1)
  • tests/models/qwen3_5/test_modeling_qwen3_5.py (modified, +2/-0)

Code Example

from transformers import AutoModelForSequenceClassification

model = AutoModelForSequenceClassification.from_pretrained(
    "Qwen/Qwen3.5-0.8B",
    num_labels=2,
    trust_remote_code=True,
)

---

ValueError: Unrecognized configuration class <class 'transformers.models.qwen3_5.configuration_qwen3_5.Qwen3_5Config'>
for this kind of AutoModel: AutoModelForSequenceClassification.

---

- Instantiate a
RAW_BUFFERClick to expand / collapse

Feature request

What happens

When trying to load a Qwen3.5 model for sequence classification:

from transformers import AutoModelForSequenceClassification

model = AutoModelForSequenceClassification.from_pretrained(
    "Qwen/Qwen3.5-0.8B",
    num_labels=2,
    trust_remote_code=True,
)

Transformers raises:

ValueError: Unrecognized configuration class <class 'transformers.models.qwen3_5.configuration_qwen3_5.Qwen3_5Config'>
for this kind of AutoModel: AutoModelForSequenceClassification.

This error occurs because Qwen3_5Config is not registered in the internal mapping of configuration classes to sequence-classification model classes. Sequence classification mappings currently include Qwen3Config and related, but not Qwen3_5Config. This prevents classification from loading via the auto class.

Expected Behavior

AutoModelForSequenceClassification.from_pretrained("Qwen/Qwen3.5-0.8B", ...) should:

  • Recognize Qwen3_5Config
  • Instantiate a Qwen3_5ForSequenceClassification class (or equivalent)
  • Allow fine-tuning and inference on classification tasks without manual patching

Metadata

Libraries / Versions

  • transformers >= 5.2.0
  • torch >= 2.x
  • Qwen3.5 model on HF Hub

Task

  • sequence classification

Motivation

Many users are adopting Qwen3.5 for fine-tuning tasks beyond generation (e.g., classification, routing), and the auto model infrastructure currently does not support classification out of the box. Addressing this will improve usability for downstream tasks.

Your contribution

I am working on implementing it and will open a PR shortly.

extent analysis

Fix Plan

Register Qwen3_5Config with AutoModelForSequenceClassification

To fix the issue, we need to register the Qwen3_5Config class with AutoModelForSequenceClassification. We can do this by creating a custom configuration class that inherits from Qwen3_5Config and registering it with the model.

Step 1: Create a custom configuration class

from transformers import AutoConfig
from transformers.models.qwen3_5.configuration_qwen3_5 import Qwen3_5Config

class Qwen3_5ForSequenceClassificationConfig(Qwen3_5Config):
    def __init__(self, **kwargs):
        super().__init__(**kwargs)
        self.num_labels = kwargs.get('num_labels', 2)

Step 2: Register the custom configuration class

from transformers import AutoModelForSequenceClassification

# Register the custom configuration class
AutoModelForSequenceClassification.register_config(
    "Qwen3_5ForSequenceClassificationConfig",
    "Qwen/Qwen3.5-0.8B"
)

Step 3: Load the model with the custom configuration class

from transformers import AutoModelForSequenceClassification

model = AutoModelForSequenceClassification.from_pretrained(
    "Qwen/Qwen3.5-0.8B",
    config=Qwen3_5ForSequenceClassificationConfig(num_labels=2),
    trust_remote_code=True,
)

Verification

To verify that the fix worked, you can try loading the model and performing a classification task:

from transformers import AutoModelForSequenceClassification
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3.5-0.8B")
model = AutoModelForSequenceClassification.from_pretrained(
    "Qwen/Qwen3.5-0.8B",

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#api#ssr#installation#tensor shape#autograd error#dependency conflict#environment setup#docker error#permission error#memory optimization

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

transformers - ✅(Solved) Fix Add AutoModelForSequenceClassification support for Qwen3.5 (Qwen3_5Config) [1 pull requests, 1 participants]