This is the Q4_K_M gguf quantized version of the GLM-4.6-REAP-266B-A32B model, built on the transformers library, with text generation capabilities. The model uses the REAP (Routing-Weighted Expert Activation Pruning) method, which can selectively remove redundant experts while retaining the router's independent control over the remaining experts.

Natural Language ProcessingTransformers

This is the Q2_K_M gguf quantized version of the GLM-4.6-REAP-266B-A32B model, created based on the REAP (Router Weighted Expert Activation Pruning) method. It can selectively remove redundant experts while retaining the router's independent control over the remaining experts, achieving efficient text generation.

Natural Language ProcessingTransformers

This is a version after quantizing Qwen3-4B-v0.4-deepresearch-no-think-4 based on llama.cpp. It is optimized specifically for in - depth research scenarios, removes the think mechanism, and uses the GGUF format to provide more efficient inference performance.

Natural Language ProcessingTransformers

This is a ViT-L14 image encoder based on the PE-Core model by Bolya et al., which has been converted to the Birder format for image feature extraction. The model retains the original weights and architecture but removes the CLIP projection layer to output raw image embeddings. It is a general-purpose visual backbone network suitable for image classification and detection tasks.

Computer VisionPytorch

This is a specific version of Voxtral-Mini-3B-2507 that removes the speech recognition function and only retains the pure language model function. This model is similar to the unreleased Ministral-3B-Instruct model and focuses on text generation tasks.

Natural Language ProcessingTransformers

Zero-Mistral-24B is an improved text-only model based on Mistral-Small-3.1-24B-Instruct-2503, primarily adapted for Russian and English, with the original visual capabilities removed to focus on text generation tasks.

Natural Language ProcessingTransformersMultiple Languages

DedistilledMixTuned Dev V3.0 is an upgraded version based on the Flux Dev fine-tuned model, achieving a high degree of balance in model capabilities. This model performs excellently in terms of LoRA compatibility, realism, image output quality, and artistic creativity, and is extremely close to the Flux Pro model. It removes anti-distillation interference through layered fusion technology and is fully compatible with the original Flux.1 Dev model.

Computer VisionDiffusersEnglish

A distilled version of the genmoai mochi-1 model transformer, composed of 42 modules (original version has 48 modules), achieving lightweight through iterative removal of modules with the smallest MSE values

MultimodalDiffusersEnglish

Mini-InternVL2-DA-RS is a multimodal model optimized for the remote sensing image domain, based on the Mini-InternVL architecture. It has been fine-tuned through a domain adaptation framework and demonstrates excellent performance in remote sensing image understanding tasks.

MultimodalTransformersOther

LibreFLUX is the Apache 2.0 open-source version of FLUX.1-schnell, restoring full T5 context length and classifier-free guidance, removing aesthetic fine-tuning content, making it more suitable for secondary development

MultimodalDiffusers

This is an ablated and quantized version based on the Google Gemma 2 27B Instruct model. It is quantized through llama.cpp, which improves the model's operating efficiency and resource utilization. This model uses innovative ablation technology to remove the rejected direction vector, making it perform better in specific scenarios.

Natural Language ProcessingTransformersMultiple Languages

A remote sensing object detection model based on the RT-DETR architecture, specifically designed for roof detection in satellite images.

Computer VisionTransformers

OpenFLUX.1 is a fine-tuned version based on the FLUX.1-schnell model, with distillation removed to allow user fine-tuning. It is a text-to-image generation model capable of producing high-quality images in 1-4 steps.

Computer VisionDiffusers

High-resolution binary image segmentation model based on bilateral reference, focusing on salient object detection and background removal tasks

Computer VisionSafetensors

BK-SDM-v2 is an efficient text-to-image synthesis model obtained by compressing SD-v2.1-base. It uses the block removal knowledge distillation technique to achieve effective distillation and retraining with limited data, providing a more lightweight solution for image generation.

Computer VisionDiffusers

An orthogonalized version of Codestral-22B-v0.1, where the model's rejection capability has been removed through ablation techniques to better align with user requests.

Natural Language ProcessingTransformersOther

A remote sensing image-specific model fine-tuned based on OpenAI CLIP, enhancing zero-shot classification and image retrieval capabilities

MultimodalJax

T5 1.1 Language Model Adaptation is an improved version of the original T5 model, using the GEGLU activation function, removing parameter sharing, and optimized specifically for language modeling tasks

Natural Language ProcessingTransformersEnglish

A remote sensing image-specific model fine-tuned based on OpenAI CLIP, enhancing zero-shot classification and cross-modal retrieval capabilities

MultimodalJax

This model is fine-tuned from bert-base-uncased-sparse-70-unstructured on the MNLI task (GLUE benchmark), with the classifier layer removed for easier loading into other downstream tasks for training.

Natural Language ProcessingTransformersEnglish