BRIEF DETAILS: LoRA model trained on FLUX.1-dev for image generation via Replicate. Uses "TOK" trigger word and integrates with 🧨 diffusers library.
Brief-details: DeepSeek V3 0324 AWQ is a quantized version of DeepSeek V3, optimized for efficient inference on high-end GPUs with impressive performance benchmarks across various configurations.
BRIEF-DETAILS: Gemma-3-Glitter-12B is a creative writing model merging RP instruction and storytelling capabilities, built on Gemma 3 12B IT with vision support.
Brief Details: Speech recognition model fine-tuned on speech commands, achieving 97.59% accuracy. Based on wav2vec2-base with excellent performance on command classification.
Brief Details: A 1.5B parameter RL-enhanced language model optimized for mathematics, achieving 49.90% average accuracy across benchmarks with GRPO algorithm
Brief Details: An 8B parameter LLaMA-based language model optimized for chat and NLP tasks. Supports multiple quantization formats and features MIT license.
BRIEF-DETAILS: Fine-tuned Whisper-small model specialized in language identification, achieving 88.6% accuracy with linear learning rate scheduling and mixed precision training
BRIEF-DETAILS: A Studio Ghibli-style LoRA adapter for FLUX.1-dev that transforms popular characters into Ghibli's distinctive artistic style, optimized for creative character illustrations.
BRIEF-DETAILS: German Text-to-Speech model featuring 4 distinct voices (Lena, Anna, Max, Tom), based on orpheus-3b-0.1-ft with ~4.5h training data per voice.
Brief-details: Zero-shot identity preservation model for ComfyUI, featuring dual-stage architecture (sim_stage1 & aes_stage2) with face preservation and aesthetics optimization capabilities.
BRIEF-DETAILS: A 12B parameter AI model focused on analytical neutrality and unbiased responses, built on Gemma architecture with specialized emotional neutrality protocols
Brief-details: Fibonacci-2-9b: A 9.24B parameter LLM based on Gemma2 architecture, optimized for NLP tasks with multi-format support (4-bit to 16-bit). MIT licensed.
Brief Details: A Gemma 3 12B model fine-tuned on Claude 3.7 Sonnet's reasoning capabilities, offering enhanced analytical thinking while remaining open-source.
Brief Details: Tessa-T1-32B is a specialized React-focused LLM based on Qwen2.5-Coder, designed for autonomous frontend development with 32B parameters and React reasoning capabilities.
Brief-details: A 7B parameter vision-language model specialized in visual mathematical reasoning and complex multimodal tasks, built on Qwen2.5-VL architecture.
Brief Details: Fibonacci-2-14B: A 14B parameter LLM based on Phi 4 architecture, optimized for NLP tasks with multi-format support (4-8-16 bit). MIT licensed.
Brief-details: DeepSeek-V3-0324 GGUF quantizations offering various compression levels from Q8_0 to IQ1_S, optimized for different hardware and memory constraints
Brief Details: TxGemma-9b-predict - A 9B parameter language model by Google requiring explicit terms acceptance via Health AI Developer Foundation
Brief-details: LHM is a groundbreaking AI model for creating animatable 3D human reconstructions from single images in seconds, featuring both 500M and 1B parameter versions
Brief-details: GenXD is an advanced AI model for generating 3D and 4D content from images, featuring multiview-temporal modules and alpha-fusing techniques. Developed by NUS and Microsoft under Apache-2.0 license.
Brief Details: Flux Wavelet model by zhang0jhon - A specialized implementation focusing on wavelet transformations for signal processing and analysis applications.
