サイト内検索

AI モデル 2119

• pegasus-large↗

  Google

  We train a pegasus model with sampled gap sentence ratios on both C4 and HugeNews, and stochastically sample important sentences. The updated the results are reported in this table. The "Mixed & Stochastic" model has the following changes: - trained on both C4 and HugeNews (dataset mixture is weighted by their number of examples). - trained for 1.5M instead of 500k (we observe slower convergence on pretraining perplexity). - the model uniformly sample a gap sentence ratio between 15% and 45%. - importance sentences are sampled using a 20% uniform noise to importance scores. - the sentencepiece tokenizer is updated to be able to encode newline character. (*) the numbers of wikihow and big_patent datasets are not comparable because of change in tokenization and data: - wikihow dataset contains newline characters which is useful for paragraph segmentation, the C4 and HugeNews model's sentencepiece tokenizer doesn't encode newline and loose this information. - we update the BigPatent dataset to preserve casing, some format cleanings are also changed, please refer to change in TFDS.

• DeepSeek-V4.1-Flash-1017-ROCMFP-GGUF↗

  Lucebox

  Lucebox builds of DeepSeek's October 17 update of V4.1 Flash (`deepseek-ai/DeepSeek-V4.1-Flash-1017`, revision `1bae0e10`) for the Lucebox DS4.1 engine (AMD Radeon AI PRO R9700 + Ryzen AI Max). Routed experts in the engine's ROCMFP block formats; attention, shared experts and head in Q8_0; both Engram tables embedded, read in place. Same recipes and the same quality harness as our September V4.1 Flash files, re-measured against the 1017 model. For reference, the September files scored KL 0.266 (ROCMFP2S), 0.246 (MIX10) and 0.204 (fp2s with fp3 down everywhere) against the September model (teacher ppl 8.45) on the same tokens. **Which one:** ROCMFP2S is what the 128 GB Lucebox installs (the hottest experts on the R9700, the rest on the Ryzen AI Max and the SSD). MIX10 is the better file where it fits resident (192 GB boards). ROCMFP23S is the best quality of the three, for two-box setups and boards with room for 179 GiB of experts. The `-IM1017` files are the same recipes calibrated on the 1017 weights; prefer them (Lucebox Manage installs `ROCMFP2S-IM1017`).

• Qwen3.8-Flash-Next-MLX-Serve-mixed-4-8bit↗

  ddalcu

  > **Use Qwen3.8-Flash-Next-MLX-Serve-iQ-MLX-4.7bpw instead.** > It is this pack re-quantized with an importance matrix collected on the bf16 > model: same layout, same 75 GB, same bytes read per token, closer to bf16 > (KLD median 0.0052 vs 0.0066, top-1 88.6% vs 87.9% against bf16's own > logits). This pack stays up for existing installs but gets no more updates. > > Switching from this pack? `ngram_table.bin` (32 GB) is identical in both. > Move it into the new pack's folder before downloading and the download > skips it: > > mlx-serve pack of Qwen/Qwen3.8-Flash-Next, the Qwen4 preview architecture (`model_type: qwen4_exp`). Runs on a 128 GB Mac with about 75 GB resident. Includes the MTP head and the vision tower (image and video input). Measured on an M4 Max 128 GB with mlx-serve 26.8.11: ~75 GB resident, decode ~60 tok/s serial and 78 tok/s with MTP (`--mtp`; +41% on code, a few percent slower on prose), prefill ~730 tok/s, a needle at 24.8k tokens recovered with sparse attention engaged. Prefix cache on, images included.

• FLUX-3-Image-Klein-GGUF↗

  AtomicChat

  Black Forest Labs' FLUX 3 Image Klein as GGUF for stable-diffusion.cpp, built from the published BF16 weights with our own importance matrix and measured on one stand against every other Klein GGUF we could find. The renders and measurements are public. Built by the team behind Atomic Chat , the app that runs open-weight models locally and offline. Everything to run Klein in stable-diffusion.cpp is here: the image model, the Qwen3-VL-4B text encoder ( text_encoders/ ) and the autoencoder ( vae/ ). The image model files also load in ComfyUI through ComfyUI-GGUF; see In ComfyUI . The prompt upsampler Klein expects is in FLUX-3-Image-Klein-pup-GGUF . AD-Q5_K stays 14% closer to the BF16 picture than unsloth's UD-Q5_K_XL at a smaller size, and AD-Q4_K 21% closer than their Q4_K_M . Where their file is better, the table below says so. The text encoder moves the picture as much as the image model does. Our AD-Q4_K encoder is 21% closer than the Q4_K_M most people use, at the same size. Looking for uncensored? The open Klein pipeline has no content filter. Our text encoder and prompt upsampler without refusals are in FLUX-3-Image-Klein-TE-Uncensored-GGUF and FLUX-3-Image-Klein-pup-Uncensored-GGUF ; what they change on the pictures is in Content filtering . FLUX 3 Image Klein is Black Forest Labs' 5.6B text-to-image and editing model: one model for generation, single-reference edits and up to ten reference images, distilled to **16 steps at CFG 1**. This repo has the image model as GGUF for stable-diffusion.cpp, with the text encoder and the autoencoder next to it.

• granite-speech-4.1-2b-plus-gguf↗

  handy-computer

  Ported from upstream commit edd3bf5, pinned 2026-05-17. Validated against the Transformers reference at transcribe.cpp commit 275332d on 2026-05-17. Offline multilingual speech-to-text with word-level timestamps. IBM Granite Speech 4.1-2b-plus is the timestamp-and-diarization variant of the Granite-Speech family. Same architecture as the base 4.1-2b (Conformer encoder, BLIP-2 Q-Former projector, Granite-4.0-1b autoregressive LLM decoder) with two changes: the encoder concatenates mid-layer (idx 3) and final-layer hidden states (`cat_hidden_layers=[3]`, doubling the projector K/V input from 1024 to 2048), and the LM token embeddings are tied with the lm_head. Takes a 16 kHz mono WAV and produces a transcript, with `--timestamps word` returning structured per-word timestamps (parsed from the model's `[T:N]` centisecond markers), or `--diarize` returning structured speaker-attributed turns from its separate SAA prompt. Those two prompt tasks cannot be combined. Transcribes English, French, German, Spanish, and Portuguese (no Japanese on this variant). This variant is transcription-only: unlike the base granite-speech-4.1-2b, it does not perform speech translation. WER on the full LibriSpeech test-clean split (2,620 utterances), batch size 1, timestamps none. Figures without a commit were published before provenance was recorded.

• glm-5.3-exlr8-k2-k3-k4↗

  kindlingai

  Version 0.5 (git tag `v0.5`). See the changelog. Weight set ID: `2011c0cedde962d19fc43e74c1d409279e398d40d27a6fdb4bcb7b49ea2ea6fb` (see Identifying the weights). Also mirrored on ModelScope: kindlingai/glm-5.3-exlr8-k2-k3-k4. > **These weights cannot be used without E4M3 rounding.** A loader must round every value the trellis > codebook produces to FP8 E4M3 (round to nearest even, saturating) before applying the side vectors. The > encoder searched only over E4M3-rounded values, so the rounding is part of the format: a stock EXL3 decoder, > or any decoder that skips the rounding, reads measurably worse weights from the same bytes.

• DeepSeek-R1-0528-Qwen3-8B-GGUF↗

  unsloth

  See our collection for all versions of R1 including GGUF, 4-bit & 16-bit formats. Unsloth Dynamic 2.0 achieves superior accuracy & outperforms other leading quants. - For Ollama do `ollama run hf.co/unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF:Q4_K_XL` - it'll auto get the correct chat template and all settings - Set the temperature between **0.5–0.7 (0.6 recommended)** to reduce repetition and incoherence. - Set Top_P value of **0.95 (recommended)** - R1-0528 uses the same chat template as the original R1 model:

• wav2vec2-large-xlsr-53-tamil↗

  Amrrs

  Fine-tuned facebook/wav2vec2-large-xlsr-53 in Tamil using the Common Voice When using this model, make sure that your speech input is sampled at 16kHz.

• SamsungSAILMontreal_ByteCraft-GGUF↗

  bartowski

• mt5-small↗

  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.