Files
HuggingFace_transformer/docs/source/en/model_doc/glm4v_moe.md
Yuxuan Zhang 7b20915f4e
Some checks failed
Secret Leaks / trufflehog (push) Has been cancelled
GLM-4.5V Model Support (#39805)
* init

* update

* uupdate

* ruff

* t patch is 2 defalut not 1

* draft

* back

* back1

* update

* config update

* update using glm-41 format

* add self.rope_scaling = config.rope_scaling

* update config

* update

* remove the processor

* update

* fix tests

* update

* for test

* update

* update 2126

* self.rope_scaling is missing in GLM4MOE lets add it

* update

* update

* Update modular_glm4v_moe.py

* change config

* update apply_multimodal_rotary_pos_emb

* format

* update

* Delete 3-rollout_qas_thinking_answers.py

* use right name

* update with place holder

* update

* use right rotary

* Update image_processing_glm4v_fast.py

* rope_config_validation needs to rewrite the entire config file in modular

* update

* changed name

* update

* Update modeling_glm4v_moe.py

* _init_weights shoud be add in Glm4vMoePreTrainedModel

* remove use_qk_norm

* Update modular_glm4v_moe.py

* remove use_qk_norm as it is not use

* fix style

* deprecations are not needed on new models

* fix merge issues

---------

Co-authored-by: raushan <raushan@huggingface.co>
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
Co-authored-by: Arthur <arthur.zucker@gmail.com>
2025-08-08 17:39:52 +02:00

2.0 KiB

PyTorch FlashAttention SDPA

Glm4vMoe

Overview

The Glm4vMoe model was proposed in by .

The abstract from the paper is the following:

Tips:

This model was contributed by [INSERT YOUR HF USERNAME HERE](https://huggingface.co/). The original code can be found here.

Glm4vMoeConfig

autodoc Glm4vMoeConfig

Glm4vMoeTextConfig

autodoc Glm4vMoeTextConfig

Glm4vMoeTextModel

autodoc Glm4vMoeTextModel - forward

Glm4vMoeModel

autodoc Glm4vMoeModel - forward

Glm4vMoeForConditionalGeneration

autodoc Glm4vMoeForConditionalGeneration - forward