Hugging Face's logo
gatilin 's Collections
Model Portfolio
ViT & Vision Encoders
Visual Tokenizers & VAEs
DiT & Visual Generation
Unified Visual Components
Scientific Multimodal Interfaces
CLIP & Embedding Experiments
Detection, Grounding & Segmentation
nanoVLM
nanoVLM Factory
Wenwu-Duobao-VL
SparkLM & GPT-2
LLM & VLM Adaptations
Hunyuan Experiments
Pending Weights
Audio Generation Components

Pending Weights

Public repositories whose audited file listings contain no recognized weight formats. Code or documentation may still be available.

  • Note Legacy ViT/vision-encoder repository with limited architecture documentation. Categorized using its name and available files; weight completeness and downstream performance were not newly validated. Pending weights: the audited listing contains no recognized weight format; inference readiness is unverified.

  • Note Legacy ViT/vision-encoder repository with limited architecture documentation. Categorized using its name and available files; weight completeness and downstream performance were not newly validated. Pending weights: the audited listing contains no recognized weight format; inference readiness is unverified.

  • Note Qwen quantized variant, with bit width identified by the repository name. Check the file listing and configuration for the actual format. This is not a standalone visual component; capabilities are not inferred from the name. Pending weights: the audited listing contains no recognized weight format; inference readiness is unverified. Model-card documentation is limited; classification uses the repository name and available configuration.

  • Note Qwen quantized variant, with bit width identified by the repository name. Check the file listing and configuration for the actual format. This is not a standalone visual component; capabilities are not inferred from the name. Pending weights: the audited listing contains no recognized weight format; inference readiness is unverified. Model-card documentation is limited; classification uses the repository name and available configuration.

  • Note SparkLM training experiments and checkpoint collection. Training configurations and versions are retained; this collection does not establish benchmark performance. Pending weights: the audited listing contains no recognized weight format; inference readiness is unverified. Model-card documentation is limited; classification uses the repository name and available configuration.

  • Note The README documents a WeDetect dual-encoder ONNX export and validation project. The Ref-4B repository name does not establish that 4B model weights are included. Pending weights: the audited listing contains no recognized weight format; inference readiness is unverified.