Moment GGUF

--- license: mit library_name: gguf pipeline_tag: time-series-forecasting language: - en base_model: AutonLab/MOMENT-1-large base_model_relation: quantized quantized_by: amaye15 tags: - gguf - time-series - forecasting - zero-shot - transformer - masked-encoder - rust inference: false --- # MOMENT-1-large (AutonLab) — GGUF GGUF conversion of [AutonLab's MOMENT-1-large](https://huggingface.co/AutonLab/MOMENT-1-large) — a masked patch encoder time-series foundation model. Converted and run with [**zsfm**](https://github.com/amaye15/zsfm-rs), a Rust workspace that ports zero-shot forecasting and tabular foundation models to GGUF + [candle](https://github.com/huggingface/candle). No PyTorch, no Python runtime required to run inference. | F32 | F16 | Q8_0 | |---|---|---| | `moment-f32.gguf` | `moment-f16.gguf` | `moment-q8.gguf` | F16 is generally the best size/accuracy trade-off; Q8_0 is smallest. This repo's default recommendation matches the upstream conversion default: **F32**. **Context must be at least 32 timesteps** (MOMENT patches the context; 32 gives a meaningful example) — a shorter context fails with `context too short`. The examples below use a 32-value context. ## Use it ### Python (`pip install zsfm`) ```bash pip install zsfm huggingface_hub ``` ```python import zsfm from huggingface_hub import hf_hub_download gguf_path = hf_hub_download("amaye15/moment-gguf", "moment-f32.gguf") model = zsfm.MomentModel(gguf_path) context = [0.85, 0.93, 1.01, 1.09, 1.17, 1.25, 1.33, 1.06, 1.14, 1.22, 1.3, 1.38, 1.46, 1.54, 1.27, 1.35, 1.43, 1.51, 1.59, 1.67, 1.75, 1.48, 1.56, 1.64, 1.72, 1.8, 1.88, 1.96, 1.69, 1.77, 1.85, 1.93] point = model.forecast(context, horizon=64) # -> List[float], the point forecast ``` ### Rust / CLI (`cargo install zsfm`) ```bash cargo install zsfm --locked ``` ```bash # downloads the original weights and converts to GGUF locally # (produces the same bytes as moment-f32.gguf in this repo): zsfm moment convert --dtype f32 -o gguf/moment-f32.gguf echo '{"context": [0.85, 0.93, 1.01, 1.09, 1.17, 1.25, 1.33, 1.06, 1.14, 1.22, 1.3, 1.38, 1.46, 1.54, 1.27, 1.35, 1.43, 1.51, 1.59, 1.67, 1.75, 1.48, 1.56, 1.64, 1.72, 1.8, 1.88, 1.96, 1.69, 1.77, 1.85, 1.93], "horizon": 64}' \ | zsfm moment infer --gguf gguf/moment-f32.gguf ``` `-m/--model` takes the full HuggingFace repo id (default `AutonLab/MOMENT-1-large`) — there's...

--- license: mit library_name: gguf pipeline_tag: time-series-forecasting language: - en base_model: moment-research/MOMENT-1-large base_model_relation: quantized quantized_by: amaye15 tags: - gguf - time-series - forecasting - zero-shot - transformer - masked-encoder - rust inference: false --- # moment-rs Pure Rust converter and inference engine for [moment-research/MOMENT-1-large](https://huggingface.co/moment-research/MOMENT-1-large). Pre-converted GGUF files are available at [amaye15/moment-gguf](https://huggingface.co/amaye15/moment-gguf). Produces GGUF v3 files and runs native forecasting — no Python required. ## Build ```bash cargo build --release ``` ## Convert Downloads the model from HuggingFace and writes a GGUF file: ```bash # F16 (recommended) ./target/release/moment-rs convert --model moment-research/MOMENT-1-large --dtype f16 --output gguf/moment-f16.gguf # Q8_0 (smallest) ./target/release/moment-rs convert --dtype q8 --output gguf/moment-q8.gguf # F32 (full precision) ./target/release/moment-rs convert --dtype f32 --output gguf/moment-f32.gguf ``` To convert all dtypes at once: ```bash ./scripts/convert_all.sh ``` HuggingFace token (optional for public models): ```bash HF_TOKEN=hf_... ./scripts/convert_all.sh ``` ## Inspect tensors Print all tensor names and shapes from a `.safetensors` checkpoint: ```bash ./target/release/moment-rs inspect-tensors models/model.safetensors ``` ## Infer Run forecasting from stdin JSON: ```bash echo '{"context": [1.0, 1.2, 1.5, 1.3, 1.8, 2.0, 1.9, 2.1], "horizon": 96}' \ | ./target/release/moment-rs infer --gguf gguf/moment-f16.gguf ``` Output is JSON in an OpenAI-compatible forecast format: ```json { "id": "forecast-000001932b7a1234", "object": "forecast", "created": 1749686400, "model": "moment", "choices": [{ "index": 0, "forecast": { "point": [2.1, 2.3, 2.5, "..."], "quantiles": {} }, "finish_reason": "stop" }], "usage": {"context_length": 8, "forecast_length": 96} } ``` **Batch / Multivariate inference** — Moment is channel-independent: each variate is encoded as an independent series. Pass a batch of univariate series to get one `Choice` per series, or use the batch mode to handle multiple variates of a multivariate dataset by submitting each variate as a separate item: ```bash # Two independent series — one Choice each echo '{"context": [[1.0, 1.2, 1.5],...

Source context: 138 downloads · 0 likes · Pipeline time-series-forecasting · Library gguf · Repo amaye15/moment-gguf