DeepSeek now trained on Ascend 950
Summary
DeepSeek has begun training its AI models on Huawei's Ascend 950 chip, building on a previous statement about advancing in AI.
Similar Articles
DeepSeek's Huawei-Chip Training Claim Finally Gets Its Benchmarks, and Its Doubters (3 minute read)
A Huawei-led consortium published benchmarks showing efficient post-training of DeepSeek's V4 models on Ascend chips, but experts note it only covers post-training and does not prove Huawei can replace Nvidia for full pretraining.
@VraserX: DeepSeek reportedly expects Huawei to start delivering chips capable of training its frontier models as early as Q4. Li…
DeepSeek expects Huawei to deliver chips for training its frontier models by Q4, with Liang Wenfeng emphasizing the necessity of this move, which could reduce China's dependency on US technology amid export controls.
Deepseek training 2T and plans 8T model
DeepSeek is training a 2 trillion-parameter AI model and plans to eventually build an 8 trillion-parameter model.
DeepSeek v4.1 Flash
DeepSeek has introduced DeepSeek-V4.1-Flash, a new AI model designed for enhanced capability, faster inference, native visual understanding, and scalability as part of their latest architecture family.
DeepSeek Introduces Vision
DeepSeek announces a new vision capability, likely a vision-language model, expanding its AI offerings.