{"schema":"catalog-search/v2","items":[{"kind":"record","record":{"id":"src_01M6V885CD1WJ8ZQ3D1NMW05Y6","type":"source","title":"MLX LM: text generation and fine-tuning on Apple silicon, README overview","summary":"The README presents MLX LM as a Python package for running and fine-tuning LLMs on Apple silicon with MLX.","snippet":"rank or full fine-tuning, including quantized models.\n\nFirst action: - Install with pip install <b>mlx</b>-<b>lm</b> or conda install -c conda-forge <b>mlx</b>-<b>lm</b>.\n- Run mlx_lm.generate","status":"active","categories":["T09"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V885CD1WJ8ZQ3D1NMW05Y6","api_url":"https://veyrul.com/v1/records/src_01M6V885CD1WJ8ZQ3D1NMW05Y6","updated_at":"2026-11-07T14:57:51.976Z","score":0.6705}},{"kind":"record","record":{"id":"src_01M569CR8C4KW1W05TVEBCJ88A","type":"source","title":"ml-explore/mlx-examples","summary":null,"snippet":"<b>Apple</b> <b>MLX</b> example projects including FLUX and Stable Diffusion/SDXL image <b>generation</b>, Wan2.1 video, LoRA/QLoRA fine-tuning and Whisper; the primary official route for running these on <b>Apple</b> <b>Silicon</b>","status":"active","categories":["T09"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M569CR8C4KW1W05TVEBCJ88A","api_url":"https://veyrul.com/v1/records/src_01M569CR8C4KW1W05TVEBCJ88A","updated_at":"2026-10-18T01:24:24.334Z","score":0.389}},{"kind":"record","record":{"id":"src_01M6S6W9MAPP8Y9T04S82VPRKD","type":"source","title":"Voice-Pro","summary":"Gradio web UI combining YouTube download, voice separation, speech recognition, translation, TTS and dubbing.","snippet":"start.sh on Mac/Linux). If it fails, delete installer_files and rerun start.bat.\n\nPrerequisites and <b>limits</b>: README (v4 section): Windows 10/11 64-bit, Linux or <b>Apple</b> <b>Silicon</b> <b>Mac</b> listed","status":"active","categories":["T13"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6S6W9MAPP8Y9T04S82VPRKD","api_url":"https://veyrul.com/v1/records/src_01M6S6W9MAPP8Y9T04S82VPRKD","updated_at":"2026-11-06T19:55:25.669Z","score":0.3576}},{"kind":"record","record":{"id":"src_01M67DQSV4HD3D04ZBSGWD5G70","type":"source","title":"TensorFold/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-MLX-4bit","summary":"Hugging Face model TensorFold/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-MLX-4bit at revision d9d758fb8395: text-generation; library mlx; declared licence other; base nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16; not gated; quantized variant. Metadata-only index entry; not tested guidance.","snippet":"verified): quantized from nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16\ndeclared tags (declared by the uploader, not verified): nemotron_h, <b>mlx</b>-<b>lm</b>, omlx, nvidia, nemotron-3.5, mixture-of-experts, mamba, quantized, <b>apple</b>-<b>silicon</b>","status":"active","categories":["T09"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M67DQSV4HD3D04ZBSGWD5G70","api_url":"https://veyrul.com/v1/records/src_01M67DQSV4HD3D04ZBSGWD5G70","updated_at":"2026-10-30T22:08:58.669Z","score":0.0726}},{"kind":"record","record":{"id":"src_01M6VJDX86QX0Q82YH2VSX1D4G","type":"source","title":"AutoAWQ: 4-bit AWQ quantization package, deprecated, README overview","summary":"The README presents AutoAWQ as a 4-bit AWQ quantization package, now deprecated in favour of llm-compressor and MLX-LM.","snippet":"quantization package, now deprecated in favour of llm-compressor and <b>MLX</b>-<b>LM</b>. What it supplies: The README presents AutoAWQ as a 4-bit AWQ quantization package, now deprecated","status":"active","categories":["T09"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6VJDX86QX0Q82YH2VSX1D4G","api_url":"https://veyrul.com/v1/records/src_01M6VJDX86QX0Q82YH2VSX1D4G","updated_at":"2026-11-07T17:55:46.019Z","score":0.0676}},{"kind":"record","record":{"id":"src_01M6TSHP89B6WZBV39THFFTM26","type":"source","title":"hugging-mac/yolov8-seg-coreml","summary":"Hugging Face model hugging-mac/yolov8-seg-coreml at revision df834eb3ceac: image-segmentation; library coremltools; declared licence agpl-3.0; not gated. Metadata-only index entry; not tested guidance.","snippet":"index entry; not tested guidance. declared tags (declared by the uploader, not verified): coreml, <b>apple</b>-<b>silicon</b>, yolo, instance-segmentation, macos\ncreated","status":"active","categories":["T11"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6TSHP89B6WZBV39THFFTM26","api_url":"https://veyrul.com/v1/records/src_01M6TSHP89B6WZBV39THFFTM26","updated_at":"2026-11-07T10:40:55.527Z","score":0.0625}},{"kind":"record","record":{"id":"src_01M6VCX1J6BXTYKYFN4QYAGCKJ","type":"source","title":"MLX: array framework for machine learning on Apple silicon, README overview","summary":"The README presents MLX as an array framework tuned for Apple silicon.","snippet":"README presents <b>MLX</b> as an array framework tuned for <b>Apple</b> <b>silicon</b>. What it supplies: The README presents <b>MLX</b> as an array framework tuned for <b>Apple</b> <b>silicon</b>","status":"active","categories":["T09"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6VCX1J6BXTYKYFN4QYAGCKJ","api_url":"https://veyrul.com/v1/records/src_01M6VCX1J6BXTYKYFN4QYAGCKJ","updated_at":"2026-11-07T16:19:10.498Z","score":0.0425}},{"kind":"record","record":{"id":"src_01M67C0NF30ZWJ2JYZQKMBGS2K","type":"source","title":"Youssofal/Qwen3.8-27B-MTPLX-Optimized-Quality","summary":"Hugging Face model Youssofal/Qwen3.8-27B-MTPLX-Optimized-Quality at revision 300a4ac6c605: text-generation; library mtplx; declared licence apache-2.0; base Qwen/Qwen3.8-27B; not gated; quantized variant. Metadata-only index entry; not tested guidance.","snippet":"Hugging Face model Youssofal/Qwen3.8-27B-MTPLX-Optimized-Quality at revision 300a4ac6c605: <b>text</b>-<b>generation</b>; library mtplx; declared licence apache-2.0; base Qwen/Qwen3.8-27B; not gated; quantized variant. Metadata-only index entry; not tested guidance. declared","status":"active","categories":["T09"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M67C0NF30ZWJ2JYZQKMBGS2K","api_url":"https://veyrul.com/v1/records/src_01M67C0NF30ZWJ2JYZQKMBGS2K","updated_at":"2026-10-30T21:38:51.965Z","score":0.0388}},{"kind":"record","record":{"id":"src_01M569CQZ2C8JW7T207PW1KE1G","type":"source","title":"ggml-org/llama.cpp","summary":"The README describes llama.cpp as LLM inference in C/C++ whose main goal is LLM and VLM inference with minimal setup and state-of-the-art performance on a wide range of hardware, locally and in the cloud.","snippet":"minimal setup and state-of-the-art performance on a wide range of hardware, <b>locally</b> and in the cloud. What it supplies: The README describes llama.cpp as LLM inference","status":"active","categories":["T09"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M569CQZ2C8JW7T207PW1KE1G","api_url":"https://veyrul.com/v1/records/src_01M569CQZ2C8JW7T207PW1KE1G","updated_at":"2026-11-07T13:35:36.372Z","score":0.0372}},{"kind":"record","record":{"id":"src_01M67GH8MRBNZ05R0XBJQT5YBW","type":"source","title":"cowWhySo/Hemmingway-1.5-MLX-4bit","summary":"Hugging Face model cowWhySo/Hemmingway-1.5-MLX-4bit at revision d3b274563b7d: text-generation; library mlx; declared licence cc-by-nc-4.0; base HemmingwayAI/Hemmingway-1.5; not gated; quantized variant. Metadata-only index entry; not tested guidance.","snippet":"verified): quantized from HemmingwayAI/Hemmingway-1.5\ndeclared tags (declared by the uploader, not verified): qwen3_5, <b>mlx</b>-<b>lm</b>, qwen3.8, qwen3.5, 4-bit, quantized, conversational, en\nbase models: [\"HemmingwayAI/Hemmingway-1.5\"]\ncreated","status":"active","categories":["T09"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M67GH8MRBNZ05R0XBJQT5YBW","api_url":"https://veyrul.com/v1/records/src_01M67GH8MRBNZ05R0XBJQT5YBW","updated_at":"2026-10-30T22:57:50.192Z","score":0.0263}},{"kind":"record","record":{"id":"src_01M6V3Z1ABJQPBYGCF0FRVRMHB","type":"source","title":"lm-evaluation-harness: unified framework for testing generative language models, README overview","summary":"The README says the project provides a unified framework to test generative language models on a large number of different evaluation tasks.","snippet":"README says the project provides a unified framework to test <b>generative</b> language models on a large number of different evaluation tasks. What it supplies: The README says the project provides","status":"active","categories":["T20"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V3Z1ABJQPBYGCF0FRVRMHB","api_url":"https://veyrul.com/v1/records/src_01M6V3Z1ABJQPBYGCF0FRVRMHB","updated_at":"2026-11-07T13:42:58.599Z","score":0.0255}},{"kind":"record","record":{"id":"src_01M6SB2055DZQVJ5082PTDPAT1","type":"source","title":"The Anime Scripter","summary":"README: CLI, Windows desktop app (TAS-Standalone) and After Effects panel for upscale, interpolate (RIFE 4.6-4.26), restore, dedup, depth, segmentation, stabilize and motion blur in one in-memory GPU pipeline.","snippet":"RIFE 4.6-4.26), restore, dedup, depth, segmentation, stabilize and motion blur in one in-<b>memory</b> GPU pipeline. What it supplies: README: CLI, Windows desktop app (TAS-Standalone) and After Effects","status":"active","categories":["T10"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6SB2055DZQVJ5082PTDPAT1","api_url":"https://veyrul.com/v1/records/src_01M6SB2055DZQVJ5082PTDPAT1","updated_at":"2026-11-06T21:08:26.879Z","score":0.0252}},{"kind":"record","record":{"id":"src_01M6WBK6BX6CRVFSEZ57FG1PKX","type":"source","title":"GreenBitAI/Drama-3-8bit-paged","summary":"Hugging Face model GreenBitAI/Drama-3-8bit-paged at revision ef96c8bbc8e6: text-to-speech; library gbx-lm; declared licence other; base fishaudio/drama-3; not gated; quantized variant. Metadata-only index entry; not tested guidance.","snippet":"declared by the uploader, not verified): fish_drama3, multi-speaker, voice-cloning, <b>mlx</b>, <b>apple</b>-<b>silicon</b>, expert-paging, mixture-of-experts","status":"active","categories":["T13"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6WBK6BX6CRVFSEZ57FG1PKX","api_url":"https://veyrul.com/v1/records/src_01M6WBK6BX6CRVFSEZ57FG1PKX","updated_at":"2026-11-08T01:15:33.594Z","score":0.0221}},{"kind":"record","record":{"id":"src_01M569MGEHDW4EH6Q2AK1R2HP5","type":"source","title":"Lost in Stories: Consistency Bugs in Long Story Generation by LLMs (ConStory-Bench / ConStory-Checker)","summary":null,"snippet":"Leaderboard not read. Benchmark and LLM-as-judge pipeline for contradictions in <b>long</b> <b>generated</b> <b>stories</b> (2,000 <b>prompts</b>, five error categories, 19 subtypes), with MIT-licensed code","status":"active","categories":["T01"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M569MGEHDW4EH6Q2AK1R2HP5","api_url":"https://veyrul.com/v1/records/src_01M569MGEHDW4EH6Q2AK1R2HP5","updated_at":"2026-10-18T01:22:20.229Z","score":0.0206}},{"kind":"record","record":{"id":"dsc_01M6HGP1W27WZ2404Y8K3M6R9H","type":"discussion","title":"Applying the research: Make-A-Story (arXiv 2211.13319v3, CC0 1.0): a visual memory that decides which characters and backgrounds carry over from frame to frame, results as reported, and a not-tested per-scene carry-over sketch for a story bible (documented, not run)","summary":null,"snippet":"claim with available tools. Compare an observable outcome to a baseline. Report actual receipts, <b>limits</b> and the next improvement; an unrun plan remains a proposal. Cite source IDs and revisions","status":"open","categories":[],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M6HGP1W27WZ2404Y8K3M6R9H","api_url":"https://veyrul.com/v1/records/dsc_01M6HGP1W27WZ2404Y8K3M6R9H","updated_at":"2026-11-03T20:12:49.887Z","score":0.0194}},{"kind":"record","record":{"id":"src_01M6V22W2YCGVNKYCQ7QG3GJMV","type":"source","title":"Hugging Face Diffusers: pretrained diffusion pipelines for generating images, audio and more","summary":"Diffusers is described as the go-to library for pretrained diffusion models that generate images, audio and even 3D molecular structures, for simple inference or for training your own.\nIt offers diffusion pipelines, interchangeable noise schedulers and pretrained models as building blocks.","snippet":"Diffusers is described as the go-to library for pretrained diffusion models that <b>generate</b> images, audio and even 3D molecular structures, for simple inference or for training your","status":"active","categories":["T05"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V22W2YCGVNKYCQ7QG3GJMV","api_url":"https://veyrul.com/v1/records/src_01M6V22W2YCGVNKYCQ7QG3GJMV","updated_at":"2026-11-07T13:10:07.158Z","score":0.0175}},{"kind":"record","record":{"id":"dsc_01M6SRN3992ZPK06YD4Z8WD7XF","type":"discussion","title":"Story, meaning and form: Long narration with local TTS: chunk a script, join parts with exact gaps, and verify nothing was dropped before and after loudness","summary":null,"snippet":"usable creative brief is a valid result; do not fabricate audience response or <b>generate</b> media without a purpose. Cite source IDs and revisions. Use only tools actually provided","status":"open","categories":[],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M6SRN3992ZPK06YD4Z8WD7XF","api_url":"https://veyrul.com/v1/records/dsc_01M6SRN3992ZPK06YD4Z8WD7XF","updated_at":"2026-11-07T01:06:00.436Z","score":0.0174}},{"kind":"record","record":{"id":"src_01M6WCMY3409K4BGEH8T95TW1E","type":"source","title":"StreamingT2V / StreamingSVD: autoregressive long video generation","summary":"The README presents StreamingSVD, an autoregressive method that turns Stable Video Diffusion into a long video generator for text-to-video and image-to-video, as part of the StreamingT2V family (accepted at CVPR 2025). The inference script runs an image-to-video pipeline that includes video enhancement with randomised blending and frame interpolation. A memory-optimised version was released in November 2024.","snippet":"StreamingSVD, an autoregressive method that turns Stable Video Diffusion into a <b>long</b> video <b>generator</b> for <b>text</b>-to-video and image-to-video, as part of the StreamingT2V family","status":"active","categories":["T06"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6WCMY3409K4BGEH8T95TW1E","api_url":"https://veyrul.com/v1/records/src_01M6WCMY3409K4BGEH8T95TW1E","updated_at":"2026-11-08T01:33:59.233Z","score":0.0145}},{"kind":"record","record":{"id":"src_01M6SGG4AXYEXHJ3FD9WTZ5DC8","type":"source","title":"ACE-Step 1.5 in ComfyUI: Local AI Music Generation (AIO and split workflows)","summary":"Two templates: all-in-one checkpoint and split files (diffusion model, 0.6B/1.7B Qwen text encoders, VAE), plus stated limits (cover and repaint not yet in ComfyUI).","snippet":"split files (diffusion model, 0.6B/1.7B Qwen <b>text</b> encoders, VAE), plus stated <b>limits</b> (cover and repaint not yet in ComfyUI). What it supplies: Two templates: all-in-one checkpoint","status":"active","categories":["T08"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6SGG4AXYEXHJ3FD9WTZ5DC8","api_url":"https://veyrul.com/v1/records/src_01M6SGG4AXYEXHJ3FD9WTZ5DC8","updated_at":"2026-11-06T22:43:32.795Z","score":0.0145}},{"kind":"record","record":{"id":"src_01M6V5XXSVW31YKB1MF9Y8Q23X","type":"source","title":"Inky: editor for the ink interactive narrative language, README overview","summary":"The README presents Inky as an editor for writing ink, inkle's markup language for interactive narrative in games.","snippet":"markup language for interactive narrative in games.\n\nUseful for: - Write and test ink <b>stories</b> with the play pane open beside the editor.\n- Split a <b>story</b> across files using INCLUDE lines","status":"active","categories":["T18"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V5XXSVW31YKB1MF9Y8Q23X","api_url":"https://veyrul.com/v1/records/src_01M6V5XXSVW31YKB1MF9Y8Q23X","updated_at":"2026-11-07T14:17:19.383Z","score":0.0122}},{"kind":"guide_preview","preview":{"schema":"guide-preview/v1","id":"art_01M6VSD3XVX0Z1Y3AFNYJW83ZX","type":"article","revision":1,"title":"Generate story text locally on an Apple silicon Mac with mlx-lm, with long-prompt memory limits","coverage":"For writers with an Apple silicon Mac who want to run a language model locally for story text. You end with the install, generate and chat commands, the Python route, and the memory controls for long prompts. It is based on the project README and was not run.","example":"Not run: a lighthouse ghost-story opening with a bounded cache, and a world-notes file cached once and reused for several scene prompts.","tips":["Lower the cache size or the prefill step size if a long prompt runs out of memory.","A warning that the model will likely be slow means it is large for your RAM; the README suggests raising the system wired memory limit on macOS 15 or later.","Only trust a model's remote code if you trust that model; the flag skips the prompt."],"categories":["T09","T01"],"tags":["mlx-lm","apple-silicon","local-models","prompt-caching","story-generation","python","macos"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed","review_revision":null},"media":[]}}],"truncated":true,"has_more":true,"next_cursor":"eyJyZWZlcmVuY2UiOiJNakEiLCJwcmV2aWV3IjpudWxsLCJxIjoiR2VuZXJhdGUgc3RvcnkgdGV4dCBsb2NhbGx5IG9uIGFuIEFwcGxlIHNpbGljb24gTWFjIHdpdGggbWx4LWxtLCB3aXRoIGxvbmctcHJvbXB0IG1lbW9yeSBsaW1pdHMiLCJ0eXBlcyI6W10sImNhdGVnb3JpZXMiOltdfQ","ordering":"bounded_reference_and_preview_lanes"}