サイト内検索

AI モデル 2119

• InternVL3-1B↗

  OpenGVLab

  We introduce InternVL3, an advanced multimodal large language model (MLLM) series that demonstrates superior overall performance. Compared to InternVL 2.5, InternVL3 exhibits superior multimodal perception and reasoning capabilities, while further extending its multimodal capabilities to encompass tool usage, GUI agents, industrial image analysis, 3D vision perception, and more. Additionally, we compare InternVL3 with Qwen2.5 Chat models, whose corresponding pre-trained base models are employed as the initialization of the langauge component in InternVL3. Benefitting from Native Multimodal Pre-Training, the InternVL3 series achieves even better overall text performance than the Qwen2.5 series. As shown in the following figure, InternVL3 retains the same model architecture as InternVL 2.5 and its predecessors, InternVL 1.5 and 2.0, following the "ViT-MLP-LLM" paradigm. In this new version, we integrate a newly incrementally pre-trained InternViT with various pre-trained LLMs, including InternLM 3 and Qwen 2.5, using a randomly initialized MLP projector. As in the previous version, we applied a pixel unshuffle operation, reducing the number of visual tokens to one-quarter of the original. Besides, we adopted a similar dynamic resolution strategy as InternVL 1.5, dividing images into tiles of 448×448 pixels. The key difference, starting from InternVL 2.0, is that we additionally introduced support for multi-image and video data.

• wav2vec2-xls-r-300m↗

  facebook

  XLS-R is Facebook AI's large-scale multilingual pretrained model for speech (the "XLM-R for Speech"). It is pretrained on 436k hours of unlabeled speech, including VoxPopuli, MLS, CommonVoice, BABEL, and VoxLingua107. It uses the wav2vec 2.0 objective, in 128 languages. When using the model make sure that your speech input is sampled at 16kHz. **Note**: This model should be fine-tuned on a downstream task, like Automatic Speech Recognition, Translation, or Classification. Check out **this blog** for more information about ASR. Authors: Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli

• Qwen3.6-35B-A3B-Uncensored-Heretic-MLX-4bit↗

  froggeric

  **1. Architecture-aware uncensoring.** Qwen 3.6 uses a hybrid attention design — linear (DeltaNet-style) *and* traditional softmax blocks, mixed 3:1. Most abliteration tools treat them the same. llmfan46 applied **separate parameters** for each attention type using the Heretic tool, yielding one of the lowest KL divergences (0.0015) of any uncensored Qwen variant — 88% fewer refusals with negligible capability loss. **2. A fixed chat template.** The official Qwen 3.6 template is broken on every C++ runtime (LM Studio, llama.cpp, MLX). Tool calls crash, the `developer` role throws errors, and empty thinking blocks waste your context window. This model ships with a rewritten template that fixes all five issues and adds a thinking toggle (` ` / ` `) you can drop into any message. **3. Vision, fixed and working.** The source model had 333 vision tower keys with incorrect prefixes, breaking image inputs. Those were corrected before conversion, so text, image, and video inputs all work out of the box.

• shap-e↗

  OpenAI

  Shap-E introduces a diffusion process that can generate a 3D image from a text prompt. It was introduced in Shap-E: Generating Conditional 3D Implicit Functions by Heewoo Jun and Alex Nichol from OpenAI. *We present Shap-E, a conditional generative model for 3D assets. Unlike recent work on 3D generative models which produce a single output representation, Shap-E directly generates the parameters of implicit functions that can be rendered as both textured meshes and neural radiance fields. We train Shap-E in two stages: first, we train an encoder that deterministically maps 3D assets into the parameters of an implicit function; second, we train a conditional diffusion model on outputs of the encoder. When trained on a large dataset of paired 3D and text data, our resulting models are capable of generating complex and diverse 3D assets in a matter of seconds. When compared to Point-E, an explicit generative model over point clouds, Shap-E converges faster and reaches comparable or better sample quality despite modeling a higher-dimensional, multi-representation output space. We release model weights, inference code, and samples at this https URL.*

• MOSS-TTS-Local-Transformer-v1.5↗

  OpenMOSS-Team

  **MOSS-TTS-Local-Transformer-v1.5** is continued from MOSS-TTS-Local-Transformer-v1.0. It preserves the main 1.0 capabilities, including zero-shot voice cloning, long-form speech generation, token-level duration control, Pinyin/IPA pronunciation control, multilingual synthesis, and code-switching. For the full 1.0 feature walkthrough, input schema, and evaluation tables, please refer to the MOSS-TTS-Local-Transformer-v1.0 README. Compared with MOSS-TTS-Local-Transformer-v1.0, v1.5 focuses on the following improvements: - **Higher-fidelity stereo audio modeling**: v1.5 uses MOSS-Audio-Tokenizer-v2 as the audio tokenizer, supporting native 48 kHz stereo input and output for richer spatial detail and more natural perceived audio quality. Since the codec output is stereo, save the `[channels, samples]` tensor returned by `processor.decode(...)` directly. - **Stronger multilingual synthesis with language tags**: when the `language` field is omitted, v1.5 may improve some languages and regress slightly on others compared with 1.0. When the language is specified, v1.5 is stronger than 1.0 on almost all supported languages. Set the tag when building the user message, for example `processor.build_user_message(text=text_fr, language="French")`. - **More stable voice cloning**: v1.5 improves speaker similarity and reduces cloning variance, making repeated generations more consistent. - **Better long-reference, short-text cloning**: v1.5 handles scenarios where the reference audio is much longer than the target text more reliably than 1.0. - **More stable punctuation-following prosody**: v1.5 follows punctuation-driven pauses more closely, especially in long sentences. - **Explicit pause control**: v1.5 supports inline pause markers such as `"[pause 3.2s]"`. For example, `我今天学习了一首中国的古诗，它的名字是[pause 3.2s]静夜思！` inserts an explicit 3.2s pause before `静夜思`. MOSS-TTS Local Transformer v1.5 supports **31 languages**. It keeps the 20 languages supported by MOSS-TTS-Local-Transformer-v1.0 and extends multilingual continued training to additional languages including Cantonese, Dutch, Finnish, Hindi, Macedonian, Malay, Romanian, Swahili, Tagalog, Thai, and Vietnamese.

• Darwin-35B-ZTC-Vision↗

  FINAL-Bench

  Darwin-35B-ZTC-Vision is a vision decision model. It does not generate text. Given one or more images, a state and a typed question, it returns a calibrated probability for every option in one forward pass: yes or no (`noul`), one of several options (`choice`), or a level on an ordered scale (`score`). It handles up to four images per request, for example "which of the two photos has the higher quality" or "is there a defect on this part". It is a 35B mixture-of-experts model (about 3B active parameters per token) built on Darwin-35B-A3B-Opus with the Darwin ZTC decision readout. It is the vision member of the ZTC series; the text member is Darwin-27B-ZTC-v9. All 20 required vision benchmarks complete (the optional Box tasks are not supported), zero failed decisions. Merged by the maintainer on 2026-10-19 (PR #76). At submission it ranked first (Borda 91.00) on the published vision results of 2026-10-17 (16 complete models) with this model added, recomputed with the leaderboard's own Borda method; that table is below. More models have entered since: on the published results of 2026-10-19 (32 complete models) the same recomputation places it fifth with 81.80. The live vision leaderboard is authoritative.

• mailguard-jev-style-1.5b↗

  JackKozmo29

  **Local System One model for enterprise inbox triage. JEV-compatible typed decisions, no API key required.** > *Drop-in local alternative to TypeSafe Jev for inbox triage pipelines. > Returns the same typed judgment schema — swap the endpoint, keep the agent.* mailguard-jev-style-1.5b is a fine-tuned Qwen2.5-1.5B-Instruct optimized to produce **typed triage judgments** over enterprise email — matching the output contract of TypeSafe System One / Jev without requiring an API subscription.

• wav2vec2-xls-r-1b-portuguese↗

  jonatasgrosman

  Fine-tuned facebook/wav2vec2-xls-r-1b on Portuguese using the train and validation splits of Common Voice 8.0, CORAA, Multilingual TEDx, and Multilingual LibriSpeech. When using this model, make sure that your speech input is sampled at 16kHz. This model has been fine-tuned by the HuggingSound tool, and thanks to the GPU credits generously given by the OVHcloud :) Using the HuggingSound library:

• nemotron-3-diarization-coreml↗

  altic-dev

  Core ML export of the final `nvidia/Nemotron-3-Diarization` checkpoint (revision `98fcee8e866bb534eee85d1eb3817170700ae717`), for on-device speaker diarization on macOS 14+ / iOS 17+. Weights are stored as fp16 and widened to fp32 at load time (`constexpr_cast`), so every op computes in fp32. The package stays at fp16 size (~199 MB). Frame-level agreement with NVIDIA's NeMo reference, same streaming loop and post-processing, on real meeting recordings:

• VoiceTut-TTS-GGUF↗

  mohammedaly22

  **Egyptian Arabic & code-switching TTS for audio.cpp: up to 6.6× faster than the PyTorch package on the same GPU, no Python** These are standalone GGUF packages of VoiceTut-TTS, an OmniVoice fine-tune trained on ~380 hours of Egyptian podcast speech. It speaks natural Egyptian Arabic, switches between Arabic and English mid-sentence (`عندي meeting بكرة`), and clones voices zero-shot. VoiceTut has exactly the same architecture as OmniVoice, so it runs through audio.cpp's existing **`omnivoice`** family with every OmniVoice route: voice cloning, voice design, streaming, the long-text chunker, and the server.