全站搜索
AI 模型 2119
- Frame_VAD_Multilingual_MarbleNet_v2.0↗
NVIDIA
Frame-VAD Multilingual MarbleNet v2.0 is a convolutional neural network for voice activity detection (VAD) that serves as the first step for Speech Recognition and Speaker Diarization. It is a frame-based model that outputs a speech probability for each 20 millisecond frame of the input audio. The model has 91.5K parameters, making it lightweight and efficient for real-time applications. To reduce false positive errors — cases where the model incorrectly detects speech when none is present — the model was trained with white noise and real-word noise perturbations. During training, the volume of audios was also varied. Additionally, the training data includes non-speech audio samples to help the model distinguish between speech and non-speech sounds (such as coughing, laughter, and breathing, etc.) **Key Features** - Lightweight model with only 91.5K parameters - Robust against false positive errors - Outputs speech probability for each 20 ms audio frame - Multilingual support: Chinese, English, French, German, Russian, and Spanish Developers, speech processing engineers, and AI researchers will use it as the first step for other speech processing models.
- mt5-base↗
Google
mT5 is pretrained on the mC4 corpus, covering 101 languages: Afrikaans, Albanian, Amharic, Arabic, Armenian, Azerbaijani, Basque, Belarusian, Bengali, Bulgarian, Burmese, Catalan, Cebuano, Chichewa, Chinese, Corsican, Czech, Danish, Dutch, English, Esperanto, Estonian, Filipino, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Haitian Creole, Hausa, Hawaiian, Hebrew, Hindi, Hmong, Hungarian, Icelandic, Igbo, Indonesian, Irish, Italian, Japanese, Javanese, Kannada, Kazakh, Khmer, Korean, Kurdish, Kyrgyz, Lao, Latin, Latvian, Lithuanian, Luxembourgish, Macedonian, Malagasy, Malay, Malayalam, Maltese, Maori, Marathi, Mongolian, Nepali, Norwegian, Pashto, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Samoan, Scottish Gaelic, Serbian, Shona, Sindhi, Sinhala, Slovak, Slovenian, Somali, Sotho, Spanish, Sundanese, Swahili, Swedish, Tajik, Tamil, Telugu, Thai, Turkish, Ukrainian, Urdu, Uzbek, Vietnamese, Welsh, West Frisian, Xhosa, Yiddish, Yoruba, Zulu. **Note**: mT5 was only pre-trained on mC4 excluding any supervised training. Therefore, this model has to be fine-tuned before it is useable on a downstream task.
- Qwen3.5-9B-NVFP4↗
ig1
This model has been quantized using llm-compressor v0.10.1.dev31+geb49917e (just after Qwen3.5 support was merged) and transformers v5.3.0. It is based on the official example with a few modifications (see next section). The sequence length has been increased from 4096 to 8192 and the number of samples from 256 to 1024. The 1024 samples come from 4 differents datasets: * 256 general conversation samples (UltraChat) * 256 math reasoning samples (GSM8K) * 256 code samples (CodeAlpaca) * 256 multilingual samples (Aya)
- Athene-V2-Chat↗
Nexusflow
We introduce Athene-V2-Chat-72B, an open-weights LLM on-par with GPT-4o across benchmarks. It is currently the best open model according to Chatbot Arena, where it beats GPT-4o-0513 (the best GPT-4o model on Arena) in hard and math category, and is on-par with GPT-4o-0513 in coding, instruction following, longer query and multi-turn. It is trained through RLHF with Qwen-2.5-72B-Instruct as base model. Athene-V2-Chat-72B excels in chat, math, and coding. Its sister model, Athene-V2-Agent-72B, surpasses GPT-4o in complex function calling and agentic applications. Athene-V2-Chat uses the same chat template as Qwen2.5-72B-Instruct. Below is an example simple usage using the Transformers library.
- Srijith-rkr_-_deepseek_base_1e-3_NO_cot_all_samples_3_epoch-gguf↗
RichardErkhov
- ThinkingCap-Qwen3.8-27B-GGUF↗
bottlecapai
GGUF / llama.cpp quantizations of bottlecapai/ThinkingCap-Qwen3.8-27B — the ThinkingCap finetune of Qwen3.8-27B that cuts reasoning tokens by 37% on average while holding 85.8% average accuracy against the base model's 86.6%. ➡️ Full model description, evaluation results (multi-seed, statistically tested), the measured thinking-token reduction, recommended sampling params, and citation: see the main model card at bottlecapai/ThinkingCap-Qwen3.8-27B. The low-bit files are built with an **importance matrix** (activation statistics from a chat-templated calibration corpus) and a **per-tensor precision layout**: the attention projections of the full-attention layers and the linear-attention output projections stay at 6–8 bit while the feed-forward weights take the 4-bit hit.
- airoboros-dpo-70b-3.3↗
jondurbin
Another experimental model, tuned primarily from synthetic data generated by airoboros, plus an additional tuning phase with various DPO datasets. This model uses the llama-3-instruct prompt template, and is provided in the tokenizer config. You can use the `apply_chat_template` method to accurate format prompts, e.g.: BEGININPUT BEGINCONTEXT [key0: value0] [key1: value1] ... other metdata ... ENDCONTEXT [insert your text blocks here] ENDINPUT [add as many other blocks, in the exact same format] BEGININSTRUCTION [insert your instruction(s). The model was tuned with single questions, paragraph format, lists, etc.] ENDINSTRUCTION
- wavlm-base-sv↗
Microsoft
The model was pretrained on 16kHz sampled speech audio with utterance and speaker contrastive loss. When using the model, make sure that your speech input is also sampled at 16kHz. The model was pre-trained on 960h of Librispeech. Authors: Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei
- seresnext26d_32x4d.bt_in1k↗
timm
This model features: * ReLU activations * 3-layer stem of 3x3 convolutions with pooling * 2x2 average pool + 1x1 convolution shortcut downsample * grouped 3x3 bottleneck convolutions * Squeeze-and-Excitation channel attention Recipe details: * Bag-of-Tricks recipe. * SGD (w/ Nesterov) optimizer * Cosine LR schedule with warmup
- Nanbeige4.3-33B↗
Nanbeige
In Nanbeige4.3, we take a foundational step toward pushing the boundaries of lightweight agents. Our strategy centers on building a high-capacity teacher agent first, which will subsequently anchor the distillation of advanced agentic capabilities into smaller models. To this end, we first introduce Nanbeige4.3-33B, designed as the flagship teacher model for the Nanbeige4.3 series. Delivering state-of-the-art agentic performance among open models in the ~30B parameter class, Nanbeige4.3-33B excels at complex coding and professional cowork tasks. It not only offers a powerful, cost-effective option for immediate real-world deployment, but also lays the necessary foundation to power our upcoming small model distillation pipeline (e.g., for Nanbeige4.3-3B). We compare Nanbeige4.3-33B with recently released frontier models at around the 30B-parameter scale: Qwen3.8-27B, Muse-Glimmer-30B, and Gemma4-31B. All models are evaluated at their highest thinking effort. For each benchmark, we use the same inference hyperparameters and harness across models.