alt.freeweights.glm / GLM-ASR-Nano-2512

From:

zai-org <huggingface.co/zai-org>Newsgroups: alt.freeweights.glmSubject:

GLM-ASR-Nano-2512

Date: 2026-11-14Commit: 61ba4e0Signed: 2DC28CC7B26A283B ✓

Automatic speech recognition model with 2.3B parameters. Safetensors weights, 4.52 GB in all. Published by zai-org under MIT.

2.3B params

4.52GBGLM-ASR-Nano-2512safetensorspermissive

FactsParams2.3BSize4.52 GBFormatssafetensorsTaskautomatic speech recognitionLicenseMITFrom sourcepermissive

$ fw get zai-org/GLM-ASR-Nano-2512
# manifest sha256 5bf52be6e53d… · 4.52 GB · 10 files
$ fw verify ./GLM-ASR-Nano-2512 zai-org/GLM-ASR-Nano-2512

From sourceHosted (signed)zai-org/GLM-ASR-Nano-2512Last check: 2026-11-14 21:26 UTC

Run at home Get help

Get help: custom fine-tune or private deploy. We review your setup before we quote.

Works with

Run23 Fine-tune7 Convert8 Scan2

Runs with

Ranked by fit to this model (safetensors), then by GitHub stars.

Transformers

Whisper LocalAI tinygrad faster-whisper WhisperX FunASR NeMo Speech Pipecat LiveKit Agents

Speech-to-Speech SpeechBrain

Show all 23

Moshi Triton Inference Server pyannote.audio RealtimeSTT ESPnet BentoML WhisperKit CTranslate2 ExLlamaV2 WhisperLive LitServe

Overview

Size: 4.52 GB
Weights: 4.52 GB (safetensors)
Parameters: 2.3B
Formats: safetensors
Task: automatic speech recognition
License: MIT (permissive)
Files: 10, each with a sha256
Publisher: zai-org
Added: 2026-11-14
Section: alt.freeweights.glm

Lineage

Original release from the publisher.

Where these bytes come from

Publisher at the pinned commit→You

Sources are tried in order. A file counts only when its size and sha256 match the manifest. Sources are checked when you ask for them, not on a timer.

Sources

Badge	Host	Last check	Result
From source	huggingface.co	2026-11-14 21:26 UTC	passed

Model card

From the publisher at commit 61ba4e0. Shown as plain text with links; images open at the publisher.

GLM-ASR-Nano-2512#

Image: image

👋 Join our WeChat community

Model Introduction#

GLM-ASR-Nano-2512 is a robust, open-source speech recognition model with 1.5B parameters. Designed for real-world complexity, it outperforms OpenAI Whisper V3 on multiple benchmarks while maintaining a compact size.

Key capabilities include:

Exceptional Dialect Support: Beyond standard Mandarin and English, the model is highly optimized for Cantonese (粤语) and other dialects, effectively bridging the gap in dialectal speech recognition.
Low-Volume Speech Robustness: Specifically trained for "Whisper/Quiet Speech" scenarios. It captures and accurately transcribes extremely low-volume audio that traditional models often miss.
SOTA Performance: Achieves the lowest average error rate (4.10) among comparable open-source models, showing significant advantages in Chinese benchmarks (Wenet Meeting, Aishell-1, etc..).

Benchmark#

We evaluated GLM-ASR-Nano against leading open-source and closed-source models. The results demonstrate that * GLM-ASR-Nano (1.5B)* achieves superior performance, particularly in challenging acoustic environments.

Image: Benchmark results

Notes:

Wenet Meeting reflects real-world meeting scenarios with noise and overlapping speech.
Aishell-1 is a standard Mandarin benchmark.

Inference#

GLM-ASR-Nano-2512 can be easily integrated using the transformers library. We will support transformers 5.x as well as inference frameworks such as vLLM and SGLang. you can check more code in Github.

Transformers 🤗#

Install transformers from source:

pip install git+https://github.com/huggingface/transformers

Basic Usage#

from transformers import AutoModelForSeq2SeqLM, AutoProcessor

processor = AutoProcessor.from_pretrained("zai-org/GLM-ASR-Nano-2512")
model = AutoModelForSeq2SeqLM.from_pretrained("zai-org/GLM-ASR-Nano-2512", dtype="auto", device_map="auto")

inputs = processor.apply_transcription_request("https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/bcn_weather.mp3")

inputs = inputs.to(model.device, dtype=model.dtype)
outputs = model.generate(**inputs, do_sample=False, max_new_tokens=500)

decoded_outputs = processor.batch_decode(outputs[:, inputs.input_ids.shape[1] :], skip_special_tokens=True)
print(decoded_outputs)

Using Audio Arrays Directly#

You can also use audio arrays directly:

from transformers import GlmAsrForConditionalGeneration, AutoProcessor
from datasets import load_dataset
from datasets import Audio

processor = AutoProcessor.from_pretrained("zai-org/GLM-ASR-Nano-2512")
model = GlmAsrForConditionalGeneration.from_pretrained("zai-org/GLM-ASR-Nano-2512", dtype="auto", device_map="auto")

# loading audio directly from dataset
ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
ds = ds.cast_column("audio", Audio(sampling_rate=processor.feature_extractor.sampling_rate))
audio_array = ds[0]["audio"]["array"]

inputs = processor.apply_transcription_request(audio_array)

inputs = inputs.to(model.device, dtype=model.dtype)
outputs = model.generate(**inputs, do_sample=False, max_new_tokens=500)

decoded_outputs = processor.batch_decode(outputs[:, inputs.input_ids.shape[1] :], skip_special_tokens=True)
print(decoded_outputs)

Batched Inference#

You can process multiple audio files at once:

from transformers import GlmAsrForConditionalGeneration, AutoProcessor

processor = AutoProcessor.from_pretrained("zai-org/GLM-ASR-Nano-2512")
model = GlmAsrForConditionalGeneration.from_pretrained("zai-org/GLM-ASR-Nano-2512", dtype="auto", device_map="auto")

inputs = processor.apply_transcription_request([
    "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/bcn_weather.mp3",
    "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/obama.mp3",
])

inputs = inputs.to(model.device, dtype=model.dtype)
outputs = model.generate(**inputs, do_sample=False, max_new_tokens=500)

decoded_outputs = processor.batch_decode(outputs[:, inputs.input_ids.shape[1] :], skip_special_tokens=True)
print(decoded_outputs)

Files

10 files · 4.52 GBWeights 4.52 GB (safetensors). Sizes and sha256 from the signed manifest.

Hide all 10 files with their sha256

File	Size	sha256
.eval_results/open_asr_leaderboard.yaml	2.13 KB	9d6b0456b243879c73c9cb30bfe632ee6feeb74a2922b35667ee1d05094ba622
.gitattributes	1.52 KB	11ad7efa24975ee4b0c3c3a38ed18737f0658a5f75a0a96787b576a78a023361
README.md	4.77 KB	9b959db4a8a0e5cbe79ecfc0939b06cd034ce69a0256de937ec516a1e58408bd
chat_template.jinja	851 B	47effbd6fb4c9dc4b0e813ed19254a3440b744f838e60cfdc47491be83a56301
config.json	1.45 KB	5e410db9ee1000a87e0b0e7bf6c5d91b313887c28f769ace68a07da66613c4c9
generation_config.json	151 B	b703d7ad51784a4b8b7a4ab95f54c166f12e18d88ceb7a8cd307591d7694af17
model.safetensors	4.52 GB	b8af83ccf6b34dfc7921cedcc46d4a6dc6aaffa661b8f71b44e3a2ff60a90a91
processor_config.json	537 B	418d55d2f3025f9428c4e23465f1fdef7ad1734f0b595602b395e76c38dadb9b
tokenizer.json	6.83 MB	6a3ea16b7964ccb586b8ef90b3044470c79a4a7b41968546347de3c04ed3b6eb
tokenizer_config.json	817 B	85b10ec24534c28dbf66e9e03a991f780d6e990aa94f352ab28b8ea8ca5831b7

VersionsEach time the pinned upstream commit or file list changed in this repo. Newest first.

Upstream commit	Date	Contents
61ba4e0b33 pinned now manifest updated, same commit	2026-11-14 21:26 UTC	10 files · 4.52 GB
61ba4e0b33	2026-11-14 21:26 UTC	10 files · 4.52 GB

Use it

$ fw get zai-org/GLM-ASR-Nano-2512
# manifest sha256 5bf52be6e53d… · 4.52 GB · 10 files
$ fw verify ./GLM-ASR-Nano-2512 zai-org/GLM-ASR-Nano-2512

fw.pyChecks the signature and every sha256, tries each source in order, resumes.

$ python3 fw.py get https://freeweights.allternit.com/manifests/zai-org/GLM-ASR-Nano-2512.json --dir ./GLM-ASR-Nano-2512
$ python3 fw.py verify ./GLM-ASR-Nano-2512 ./zai-org__GLM-ASR-Nano-2512.json

hfThe Hugging Face CLI with one setting. Same repo id.

$ HF_ENDPOINT=https://freeweights.allternit.com hf download zai-org/GLM-ASR-Nano-2512 --revision 61ba4e0b3309b6656edea3e93e419f7bd5c61957 --local-dir ./GLM-ASR-Nano-2512

transformersSpeech recognition with the transformers pipeline.

import os
os.environ["HF_ENDPOINT"] = "https://freeweights.allternit.com"
from transformers import pipeline
asr = pipeline("automatic-speech-recognition", model="zai-org/GLM-ASR-Nano-2512", revision="61ba4e0b3309b6656edea3e93e419f7bd5c61957")
print(asr("audio.wav")["text"])

Run at home

License

Name: MIT License
Id: MIT
Class: Permissive
Redistribution: yes
Listing: Hosted (signed)
Decided: License as declared by the publisher. No person read it. 2026-11-14
Text in: README.md
Text sha256: 9b959db4a8a0e5cbe79ecfc0939b06cd034ce69a0256de937ec516a1e58408bd
Terms gate: No extra terms beyond the license.

Read the license text

Licenses are checked against the license table. A custom license passes only after a person reads it.

Provenance

From:

zai-org <huggingface.co/zai-org>Newsgroups: alt.freeweights.glmSubject: GLM-ASR-Nano-2512Date: 2026-11-14Commit: 61ba4e0Signed: 2DC28CC7B26A283B ✓

Upstream: zai-org/GLM-ASR-Nano-2512
Pinned commit: 61ba4e0b3309b6656edea3e93e419f7bd5c61957
Manifest: GLM-ASR-Nano-2512.json · .minisig
Manifest sha256: 5bf52be6e53d578b2b68eea2b05929a3a96456b609001cbb6c734107c3236757
Signature: minisign, Ed25519, key 2DC28CC7B26A283B
Created: 2026-11-14 21:26 UTC
Generator: freeweights-manifest/0.1.0
Takedowns: None.

How to verify

$ curl -O https://freeweights.allternit.com/manifests/zai-org/GLM-ASR-Nano-2512.json
$ curl -O https://freeweights.allternit.com/manifests/zai-org/GLM-ASR-Nano-2512.json.minisig
$ minisign -Vm GLM-ASR-Nano-2512.json -P RWQ7KGqyx4zCLXtUFOqvfv8L3g+8fpKJ1FsuxPyOYMF+4KufbUxKq1Oq
$ shasum -a 256 <file>   # compare with the Files tab

A manifest whose signature fails must be rejected. A file whose size or sha256 does not match must be discarded. Verify a download

Model tree

No related models in this catalog.Nothing here lists this model as its parent, and it names no parent.

Get it

Download

$ HF_ENDPOINT=https://freeweights.allternit.com hf download zai-org/GLM-ASR-Nano-2512

Open in Freeweights Studio

Pull, verify and run it in the Allternit app on Desktop, web or phone.

Signed manifest

minisign, key 2DC28CC7B26A283B. GLM-ASR-Nano-2512.json. Provenance. Takedown history: none.

Source

huggingface.co/zai-org/GLM-ASR-Nano-2512

Related models

All of alt.freeweights.glm →

From: bartowski <huggingface.co/bartowski>Subject: arcee-ai_GLM-4-32B-Base-32K-GGUF

From:

ggml-org <huggingface.co/ggml-org>Subject: AutoGLM-Phone-9B-GGUF

From:

zai-org <huggingface.co/zai-org>Subject: cogagent-chat-hf

From:

zai-org <huggingface.co/zai-org>Subject: CogView3-Plus-3B

From:

zai-org <huggingface.co/zai-org>Subject: cogvlm-base-490-hf

From: laion <huggingface.co/laion>Subject: dev_set_part1_10k_glm_4_7_traces_jupiter

GLM-ASR-Nano-2512

Works with

Runs with

Fine-tune with

Convert with

Scan with

Overview

Model card

GLM-ASR-Nano-2512#

Model Introduction#

Benchmark#

Inference#

Transformers 🤗#

Basic Usage#

Using Audio Arrays Directly#

Batched Inference#

Files

Use it

License

Provenance

How to verify

Model tree

Related models