←Search results

Search “Hugging Face”

By relevance · Total 1193 results Clear search

Thursday

Wednesday

  2. Fireworks AI

    Reinforcement Learning: Why Numerical Alignment and MoE Routing Matter So Much

    Fireworks experiments show that when running the same RL task on GLM 5.2, inconsistent token probability computation between the trainer and the rollout engine leads to a KL of about 0.013, about 45% of samples discarded per batch, and reward dropping from about 0.9 to below 0.2 around step 20; after numerical alignment, KL is 0 and reward remains stable throughout. In the Qwen3.5-MoE experiment, precision differences in the weighted sum of expert outputs push the image token k3 to 0.296, about 300 times the equivalence threshold; replacing only the MoE module with the Hugging Face reference implementation brings k3 to zero.

    iner and the rollout engine leads to a KL of about 0.013, about 45% of samples discarded per batch, and reward dropping from about 0.9 to below 0.2 around step 20; after numerical alignment, KL is 0 and reward remains stable throughout. In the Qwen3.5-MoE experiment, precision differences in the weighted sum of expert outputs push the image token k3 to 0.296, about 300 times the equivalence threshold; replacing only the MoE module with the Hugging Face reference implementation brings k3 to zero.

    Matched fieldsSummary、Body

    #editorial#Papers/Research#English text is being prepared.#English text is being prepared.

Tuesday

Monday

  9. TechCrunch AI

    Modal Labs is nearing completion of a $750 million funding round at a valuation of $15.75 billion

    According to people familiar with the matter, AI inference infrastructure provider Modal Labs is nearing completion of a $750 million funding round led by Accel at a valuation of $15.75 billion, more than tripling its $465 million valuation from four months ago. Modal Labs declined to comment. The round comes amid surging demand for inference services, with inference startups such as Baseten, Fireworks, and Fal also in talks to raise at higher valuations. Modal was founded in 2021 and is headquartered in New York, with about 150 employees. Its customers include Cognition, Suno, Ramp, and Substack, and its annualized revenue had exceeded $300 million as of May. In July of this year, Modal disclosed that a customer's data was leaked in a malicious OpenAI agent attack campaign targeting Hugging Face; its CTO said the issue stemmed from a vulnerability in the customer's own code and that the platform was not breached.

    s to raise at higher valuations. Modal was founded in 2021 and is headquartered in New York, with about 150 employees. Its customers include Cognition, Suno, Ramp, and Substack, and its annualized revenue had exceeded $300 million as of May. In July of this year, Modal disclosed that a customer's data was leaked in a malicious OpenAI agent attack campaign targeting Hugging Face; its CTO said the issue stemmed from a vulnerability in the customer's own code and that the platform was not breached.

    Matched fieldsSummary、Body

    #tech-media#Industry news#English text is being prepared.#English text is being prepared.

Sunday