[KongchangAI]
ModelMNN-Pocket

MNN-Pocket-0.3B

An on-device small-parameter large language model released by Alibaba's MNN team, employing a 300M-parameter dense architecture and supporting INT4 quantization and long context, designed for deployment on memory-constrained devices.

Timeline (last 90 days)

Nov 10

MNN-Pocket-0.3B is a 300M-parameter dense model.

Unverified50%
Nov 10

MNN-Pocket-0.3B has a deployment size of 173MB under INT4 quantization.

Unverified50%
Nov 10

MNN-Pocket-0.3B achieves a maximum decoding throughput of 466 tokens/s on an iPad Pro M5 (Metal backend).

Unverified50%
Nov 10

MNN-Pocket-0.3B supports a context length of up to 32K tokens.

Unverified50%
Nov 10

Both the model weights and the MNN deployment package of MNN-Pocket-0.3B are open-sourced.

Unverified50%
Nov 10

MNN-Pocket-0.3B was released by the Alibaba MNN team.

Unverified50%
Nov 10

The pretraining of MNN-Pocket-0.3B covers approximately 4.1 trillion high-quality Chinese and English tokens.

Unverified50%
Nov 10

The pretraining of MNN-Pocket-0.3B uses a four-stage pipeline.

Unverified50%
Nov 10

The post-training of MNN-Pocket-0.3B consists of three stages: instruction fine-tuning, multi-teacher on-policy distillation, and quantization-aware training (QAT).

Unverified50%
Nov 10

MNN-Pocket-0.3B adopts dependency-aware vocabulary pruning in its architecture design.

Unverified50%

2 more timeline events

All Facts (12)

Source Articles