Instructions to use maxiboch/zipformer-streaming-ctc-small-en with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- K2
How to use maxiboch/zipformer-streaming-ctc-small-en with K2:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
zipformer-streaming-ctc-small-en
Streaming Zipformer2 encoder with a CTC head, trained on LibriSpeech by the
k2-fsa / icefall project (--causal 1 --use-ctc 1, chunk sizes 16/32/64 at
50 Hz, left context 64/128/256), repackaged as safetensors for the
maxi-ml native port
(asr-models::zipformer).
- Source checkpoint:
csukuangfj/icefall-asr-librispeech-streaming-zipformer-small-2024-03-18,exp-ctc-rnnt-small, weights averaged over epochs 28–30 exactly as the upstreamctc-epoch-30-avg-3-chunk-16-left-128.onnxexport (average_checkpoints_with_averaged_model(epoch-27.pt, epoch-30.pt)). - Converter:
scripts/convert_zipformer_checkpoint.pyin maxi-ml. Only the inference parameters are kept (encoder_embed.*,encoder.*,ctc_output.*); the transducer decoder/joiner are dropped. - Files:
model.safetensors(fp32, 22.46 M parameters),config.json(architecture hyper-parameters),tokens.txt(BPE 500 +<blk>). - Features: 80-bin Kaldi fbank, 16 kHz, dither 0, snip_edges false, high_freq −400.
- CoreML packages:
coreml/zipformer-chunk{16,8}-left128.mlpackage: the icefall streaming forward traced with coremltools (scripts/coreml/convert_zipformer_to_coreml.py), caches inMLState. Runs on the GPU; the ANE compiler rejects it.coreml-ane/zipformer-chunk{16,8}-left128.mlpackageand…-stateless.mlpackage: the same step emitted directly as a MIL program in the(B, C, 1, S)layout the ANE accepts (maxi-ml convert-zipformer-ane,maxi_ml_compile_mil::zipformer_emitter, maxi-ml commit bc970670e), caches packed into 14MLStatetensors, or carried as inputs/outputs in the-statelesstwin for macOS 27.0, which cannot loadMLStateprograms. Both have the traced package's I/O (x: [1, 2·chunk+13, 80]fp32 in,log_probs: [1, chunk/2, 500]fp32 out; the stateless one adds<cache>inputs and<cache>_outoutputs). Measured withexamples/zipformer_latency.rs: 1.6–2.5 ms per 320 ms chunk on the ANE across M3 / M4 Pro / M5 Max, 4–6 ms at microphone pacing, same WER as the fp32 MLX port on the corpus (one near-tie word can differ at chunk 16 — fp16 end to end).
- License: Apache-2.0, as upstream.
- Downloads last month
- -
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support