搜索“DeepSeek”

ormance comparison between Day-0 and v0.5.15.post1 on 8 B300 Background GLM-5.2 keeps the same backbone as earlier GLM checkpoints: DSA with a sparse-attention indexer on top of a DeepSeek-V3-style MoE. It adds two major architectural changes: IndexShare for DSA, and MTP with IndexShare and KVShare. SGLang has supported the GLM-5.2-NVFP4 checkpoint since day 0 on (Grace) Blackwell hardware, using trtllm-gen kernels for both sparse attention and MoE. To turn that day-0 stack into a faster, more s

匹配字段正文

核心的经济代言人来代表针对它们的行动可能带来的负面影响。从近期围绕Fable（以及随后的GPT-5.6）模型许可协议相关事件的报道来看，关于6月9日发生了什么，有更多说法： 据一位了解该会议的人士称，会上提到了该项目将如何处理开源AI模型的话题。该人士说，美国开源模型提供商Reflection AI的一位代表主张，开源模型应基于其能力获得该框架的豁免。目前，DeepSeek等中国开源模型相比其他可用的开放模型拥有相当大的领先优势，而Reflection尚未推出公开模型。 在这里实施任何形式的禁令，对AI的长期发展轨迹都将是一个重大错误。 最可能即将采取的行动是，禁止或无限期推迟任何能力明显高于GPT 5.5、Claude Opus 4.8或GLM-5.2区间的开放权重模型。鉴于持续存在的能力差距，这应该会在未来6个月内发生。 就目前情况而言，这些模型最有可能来自一家中国公司，正因如此，这场关于前沿开放模型能力的讨论不可避免地与其他问题（如蒸馏）联系在一起。政府“审查权”的能力门槛会随时间推移而变化，但一旦确立，对开放模型的推进很可能远慢于对闭源模型。部分原因是闭源模型更容易保障安全，但

匹配字段正文

the drive to solve very difficult problems once reserved for proprietary models. Ollama’s cloud is the most delightful way to access the most powerful open models: GLM, Nemotron, DeepSeek, Kimi, MiniMax and more. On average, Ollama’s cloud has more than doubled in token volume every month, and we can’t wait to make it available to every team. Today, we’re announcing that Ollama has raised $88M from Peter Fenton at Benchmark, Tomasz Tunguz at Theory Ventures, Alex Kolicich at 8VC, along with Sol

匹配字段正文

naturally we want to train our own model on longer horizon tasks. This introduces two challenges. First, rollouts can extend far beyond the raw context window. Second, as shown by DeepSeek R114, RL on reasoning tasks tends to produce progressively longer responses, but we want the model to be efficient in its reasoning and only elaborate on difficult tasks. We address these issues with training for self-compaction and an alternating length penalty. When an agent approaches the context limit, we

匹配字段正文

on coding agents. Developing new models from open-source starting points, however, presents a distinct challenge. Many of the strongest open-source models, such as Kimi K2.7 Code, DeepSeek-V4, and GLM 5.2, come from AI labs based in mainland China. Models from these labs raise two concerns. First, studies have found these models often repeat Chinese Communist Party ("CCP") aligned narratives considerably more than their American counterparts2,7. This is perhaps unsurprising, because these labs a

匹配字段正文

platform for building and operating agentic AI applications. Foundry starts with the widest model selection on any cloud — models from Microsoft, OpenAI, Anthropic, Meta, Mistral, DeepSeek, Hugging Face, and others, spanning frontier, open-source, and custom weights — all accessible through a single endpoint and a single set of SDKs in Python, C , JavaScript, and Java. On top of those models sits the Foundry Agent Service : multi-agent orchestration with built-in memory, knowledge grounding thro

匹配字段正文

w the multi-token prediction (MTP) story. Autoregressive models generate one token per forward pass, and that’s expensive. So people bolt on lightweight heads, à la Medusa, EAGLE, DeepSeek’s MTP, that peek at the backbone’s hidden states and guess the next few tokens in one forward pass. Pair that with speculative verification, and we get real speedups. It’s a beautiful idea, and it works. So naturally we asked: can we make it work beyond autoregressive models? Diffusion LMs (DLMs) are a natural

匹配字段正文

a systems, it’s exploring all possible logical consequences of any kind of axiom. Dwarkesh Patel On the point about whether you can provide process-based supervision without Lean, DeepSeek had their DeepSeek Math model. They released a paper on how they trained it, and it was quite interesting. The problem with natural language proofs is you don’t know if it’s correct or not. They have a verifier, and the verifier is trained by a meta-verifier that makes sure that for all the problems they’re tr

匹配字段正文

t. Where the frontier labs go Seventy is what disciplined engineering gets you. Going further means changing the architecture itself, and here the labs are placing different bets. DeepSeek’s multi-head latent attention squeezes each token’s key-value pair into a small latent rather than storing it in full, which both speeds inference and stretches context toward a million tokens; it was one of the more copied ideas after DeepSeek-R1. Qwen swaps standard quadratic attention for a linear network o

匹配字段正文

es: “Pure” model makers : These are companies whose stated goal is to train models that are at the frontier, or at least close to it. This includes many Chinese companies, such as DeepSeek, Zhipu, and Minimax, but also Western ones like Poolside, Arcee, and Zyphra. It also increasingly includes sovereign AI players, such as Cohere, Sovereign, Mistral, and Trillion Labs. The recent Mythos episode has woken up some policymakers, which may lead to increased interest in sovereign model training. Big

匹配字段正文

trained models and then train them further (”mid-training”) to apply them to our specific task. "Pre-training" — Starting from generic draft models Building off of the work of the DeepSeek team, many models are now released with multi-token prediction (MTP) heads that can be used for speculative decoding. These heads improve quality in training and performance in inference, so they’re something of a no-brainer. But for that same reason, they are first-and-foremost a train-time optimization. Bett

匹配字段正文

之一，但那是另一个话题了。有一个在社区中评价褒贬不一的基准测试（尤其是在真正的设计师群体中）——Design Arena 甚至显示 GLM-5.2 击败了 Claude Fable 本身——那个最近被封禁的炒作机器！ 在 AI 评论界和研究界中，我尊重的几乎所有人都在亲自使用后称赞了这款模型。社区中如此集中的讨论焦点，此前只有一次在开源模型发布时如此明显——DeepSeek R1。这不是我轻易做出的比较，而当我将 Kimi K2 的发布比作“DeepSeek 时刻”时，GLM-5.2 已经远远超过了那一次。Kimi K2 令人印象深刻之处在于，开源模型性能的大幅进步似乎可以来自中国的任何地方。而 GLM-5.2 所迈出的这一步，更像是 AI 进步的一扇单向门。 Anthropic 凭借 Claude Code 创下的创纪录营收增速，很大程度上是由其作为最佳模型、以及唯一真正能做到这一点的模型所驱动的。GLM-5.2 是众多（即将到来的）开源权重模型中第一个提供可信替代方案的。这一类比非常清晰，就像当年 DeepSeek R1 表明，资源远为匮乏的开源权重实验室也能复现 OpenAI 以

匹配字段正文

and backward. FP8 training As noted earlier, trainer↔inference mismatch can hurt your training. To fight this, we use DeepGEMM kernels to perform block-scaled FP8, as proposed by DeepSeek V3. This, contrary to popular belief, doesn't really increase the throughput, due to quantization overhead (except for specific configurations); however, it heavily decreases KL mismatch between trainer and inference, as both now use the same precision, and even the same kernels in some cases. This in turn lea

匹配字段正文

ed. You can try that out here. We walk through a sample output from this model below. The chart compares the decode speedup from speculation, over the autoregressive baseline, for DeepSeek-V4 Pro deployed on a single B200 node processing a specific workload. The workload in this case is short sequence length ( 4k tokens/sequence) and high concurrency (batches of 32 sequences). The chart at the top shows the predicted speedup factor at a variety of different draft lengths for several drafters, al

匹配字段正文

sual generative AI. Our team has co-developed 3 of the top 5 most popular open generative AI models of all time, including Stable Diffusion, and the only open models as popular as DeepSeek. We’ve talked a lot about language models today. They are remarkable. But language is a compressed, and ultimately limited, representation of the world. We are building models that learn from thousands of years of video, and can reason in complex visual environments—the rich, messy, complicated world we actual

匹配字段正文

[1] Ling-2.6-1T model card [2] Hybrid models meet SGLang (blog) [3] Ragged Paged Attention [4] Fused MoE V1 kernel, tpu-inference [5] Fused MoE V1 kernel adapted in SGLang-JAX [6] DeepSeek-V2 (MLA) [7] Gated Linear Attention (GLA) [8] MiniMax-01 (Lightning Attention) [9] SGLang (RadixAttention) Acknowledgments SGLang-JAX team: jimoosciuc, Prayer, aolemila, neo, leos, pathfinder-pf, Haolin Fu, Qinghan Chen, JamesBrianD, Haoguang Cai, Yuhao Hu, cjx0709, Zhengke Zhou, Yuxin Wei, Lianfang Wang, 0xas

匹配字段正文

a good idea. If the Tensordyne architecture pans out and can support all of the big inference engines, and Tensordyne can get enough HBM to make them in high volume, this may be a DeepSeek moment for AI hardware. RK Anand, one of the co-founders of Tensordyne, tells The Next Platform that Broadcom is the chip shepherd for the Napier compute engine and that with Broadcom being the third largest buyer of HBM memory on the planet and the third largest buyer of chip wafers from Taiwan Semiconductor

匹配字段正文

no impact on model quality. However, many speculative decoding methods, like the EAGLE series and the native multi-token prediction (MTP) modules in recent models like Gemma 4 and DeepSeek-V4, still rely on sequential autoregression – but in the draft model instead of the target. The draft model generates draft tokens one-by-one,a poor fit for modern hardware and a limit on achievable speedup. That’s why Z Lab developed DFlash, which uses a lightweight block diffusion draft model to generate an

匹配字段正文

nd so are harder to quantize. Models like gpt-oss combine single-precision attention operations with 4 bit matmuls to get the best of both worlds. However, key model families like DeepSeek-V3 and V4 natively (i.e., from training) support 8 bit attention operations. And other models like the Qwen and Gemma series are sometimes deployed with 8 bit KV caches to accelerate inference. So we added support for 8 bit floats (with either four or five exponent bits, aka e4m3 or e5m2). Relative to the othe

匹配字段正文

minute task timeout. Tasks took 6m 53s on average. That makes Fable slower per task than GPT 5.5, Opus 4.7, and BU 2.0, but faster than Gemini 3 Pro, Qwen 3.6 Plus, Kimi K2.6, and DeepSeek V4 Pro. The cost comparison makes the tradeoff clear. Fable completed many of the long research and browsing tasks that usually separate stronger web agents from weaker ones. These are tasks that require following many steps of reasoning, checking constraints, and returning a structured answer rather than extr

匹配字段正文