Brief Details: CodeActAgent-Mistral-7b is an LLM agent that uses executable Python code actions, built on Mistral-7B with 32k context window, optimized for tool use and general conversation.
BRIEF DETAILS: Cross-lingual Persian-English LLM achieving SOTA results on Belebele benchmark & ParsiNLU tasks. Developed by University of Tehran researchers. Size: 13.7GB (BF16).
Brief Details: 4x upscaling ESRGAN model optimized for JPEG compression, trained on diverse datasets for 150k iterations. Universal application with strong detail generation.
Body language detection system combining MediaPipe and OpenCV with dual model architecture (Scikit-learn and TensorFlow-Keras) for accurate emotion and gesture recognition across 10 distinct categories.
BRIEF-DETAILS: Stanford healthcare AI model requiring CITI certification. Specialized for EHR data analysis with strict ethical usage requirements.
Brief-details: A variant of the LLaVA (Large Language and Vision Assistant) model architecture, optimized for lightweight deployment with random initialization
Brief-details: A compact variant of the Aquila2 language model developed by katuni4ka, focused on efficient processing while maintaining core functionalities.
Brief Details: A compact chat model by katuni4ka available on HuggingFace, designed for lightweight conversational AI applications
Brief Details: A compact variant of the XVERSE model architecture by katuni4ka, focusing on efficiency while maintaining core capabilities through randomized parameters.
Brief Details: A compact random initialization variant of InternLM2, created by katuni4ka, offering a baseline experimental model for studying transformer architectures.
BRIEF-DETAILS: Advanced instance segmentation model by Facebook using masked-attention Transformer architecture. Optimized for COCO dataset with Swin-tiny backbone.
Brief-details: A lightweight CPM-based language model by katuni4ka, featuring minimalistic architecture. Available on HuggingFace, suitable for experimental and research purposes.
Brief-details: A 1.5B parameter distilled version of DeepSeek-R1, optimized for 4-bit quantization. Offers efficient reasoning capabilities while maintaining low memory footprint.
BRIEF-DETAILS: 4-bit quantized version of Microsoft's Phi-3.5-mini-instruct model optimized for MLX framework, enabling efficient deployment on Apple Silicon
Brief-details: A specialized model for detecting overlapping speech segments in audio recordings, helping identify when multiple speakers talk simultaneously
Brief Details: Mixtral-8x22B-Instruct-v0.1 is a powerful instruction-tuned language model from Mistral AI featuring 8-expert MoE architecture with 176B total parameters.
Brief Details: KoBERT - Korean BERT model optimized for Korean language processing, requiring trust_remote_code for tokenizer implementation. By monologg.
BRIEF-DETAILS: BERT base Japanese model using character-level tokenization with whole word masking, trained on Wikipedia data. Features 12 layers, 768-dim hidden states, 12 attention heads.
BRIEF-DETAILS: Meta's 405B parameter LLaMA variant optimized for instruction-following, implementing FP8 quantization for improved efficiency
Brief Details: AsianModel by BanKaiPls - A specialized AI model hosted on HuggingFace focusing on Asian-specific content processing and generation.
BRIEF-DETAILS: DeepSeek LLM 7B Base - Advanced language model with 7B parameters, trained on 2T tokens. Supports English/Chinese, open-source, MIT licensed.
