AITOPAI热报
Models

Underdog Bark 3.0：仅 163 MB 的语音转文字模型

Gorden Sun@Gorden_Sun
Post
View the post
Why we picked it

163 MB 的语音转写模型，M 系列芯片上半秒转完 5 分钟音频，错字率还比 Whisper 低，Clean 版连“嗯啊”都帮你删了。

基于英伟达与 Moondream 开源的语音识别架构，Underdog Bark 3.0 被压缩到 163 MB。在苹果 M 系列芯片上转写 5 分钟音频只需约半秒，英文会议场景下的词错率低于 Whisper 等传统模型。其 Clean 版本可自动过滤“嗯”“啊”和口误重复，直接输出干净的文本。目前仅支持英文。

Original · Gorden Sun

Underdog Bark 3.0：仅163 MB的语音转文字模型

基于英伟达与 Moondream 开源的语音识别架构，模型体积被压缩到了 163 MB，转写速度极快，在苹果 M 系列芯片上转写 5 分钟音频仅需半秒。它不仅在英文会议场景下的错误率明显低于传统的 Whisper 等大模型，其特有的 Clean 版本还能在打字时自动过滤掉日常讲话中的“嗯”、“啊”、口误重复等废话，直接输出干净顺畅的最终文本。

可惜仅支持英文。

模型：https://t.co/h43fVDIGiN

#Underdog Bark 3.0#Whisper#Moondream#语音识别#开源模型

Keep reading

Recent stories with the same tag or section
ModelsBuildersSelected17:26
vLLM Ships Day-0 Support for Sage-A1.5

Day-0 support means the model can be self-hosted on existing vLLM infrastructure immediately, without waiting for community integrations. With only 3B active parameters, it targets teams running agentic workloads on modest local hardware who want an alternative to larger dense models. Developers comparing open-source MoE options for on-premise serving should evaluate its throughput against Qwen and DeepSeek MoE checkpoints.

#Sage-A1.5#vLLM#Shanghai AI Lab#open-source models
官方账号vLLM@vllm_projectSource
ModelsBuildersSelected17:09
Tencent Hunyuan releases WorkForge agent training framework

Training agents on verifiable, evidence-grounded tasks addresses the core bottleneck of RL for long-horizon work: reward reliability. The 16.7K-environment scale makes this one of the larger open efforts in agent environment construction. Engineers building work or office agents can study the dual-verifier design for their own training pipelines.

#Tencent Hunyuan#WorkForge#agents#AI agents
官方账号腾讯混元@TencentHunyuanSource
ModelsBuildersMulti-sourceSelected16:36
Muse Glimmer 30B: Meta's distilled agent model runs on one consumer GPU

Distilling from a large teacher into a small model specialised for long tool-use chains may matter more than raw size for local agents, a trade-off the OSWorld gap makes visible. Builders running agents on local hardware can use it as a drop-in for multi-tool MCP workflows without cloud dependency. Compared with Qwen3.6-27B, it wins on tool orchestration but loses on general OS-level tasks.

#Muse Glimmer#Meta#open-source models#agents
Event
rohanpaul_ai@rohanpaul_ai4 sourcesSource