Register and share your invite link to earn from video plays and referrals.

Search results for 柯基臀
柯基臀 community
One keyword maps to one global community path.
Create community
People
Not Found
Tweets including 柯基臀
DeepSeek V4.1 Gives Prefill and Decode Different Compute Paths Prompt tokens mostly traverse 20 layers; generated tokens traverse all 40. This Causal Encoder-Decoder design nearly halves long-input prefill while preserving autoregressive generation. Zhihu contributor 潜龙勿用, Changxin Ke(柯昌鑫), a graduate researcher at ICT, CAS, explains how its architecture and post-training were designed together. 1️⃣ A causal encoder, not T5 The 40-layer backbone is split into a 20-layer causal encoder and a 20-layer decoder. Both remain causal. The encoder processes the prompt and supplies the decoder’s global KV. Most prompt tokens avoid the decoder stack, while generated tokens still run through all 40 layers. That means 8B active parameters per prefill token versus 16B per decode token. 2️⃣ Most layers share global memory CSA2 uses three modes: Full creates global KV and an index; Reindex shares the KV but selects new positions; Reuse shares both. Only four layers create independent global KV, four reindex it, and 30 reuse both KV and the latest index. Each layer retains its own query and local SWA state. The first Full layer builds up to 16,384 candidates. Later layers search this pool for their Top-512. From 4K to 1M context, decode FLOPs per token rise only about 25%. 3️⃣ Serving approximations enter training Exact reconstruction of decoder SWA states would replay 2,560 prompt tokens. V4.1 replays only the final 128 encoder outputs and trains the model to tolerate the approximation. With FP4 global KV, cache falls to 890 bytes per token, persistent cache to roughly one eighth of V4-Flash, and prefill compute close to half. Bounded replay and constrained retrieval are not last-minute serving tricks. The model experiences them during post-training. 4️⃣ Post-training is an evolving Agent system Each RL task combines a problem, environment, and verifier. New trajectories can reveal shortcuts, broken environments, or verifier errors and send the task back for repair. V4.1 trains across multiple harnesses, merges checkpoints between RL runs, and finishes with on-policy distillation from more than 40 teachers. Raising reasoning effort from 25 to 100 increases output length about 2.5×, while average Pass@1 across eight benchmarks rises from 67.1% to 76.3%. ✅ The real design choice V4.1 aligns model structure, cache policy, retrieval limits, training environments, and inference around long-running Agents. The model learns under the same constraints the deployed system will actually impose. 🔗 Full analysis: #DeepSeek# #DeepSeekV41# #LLMArchitecture# #AIAgents# #LongContext# #AIInfra#
Show more
You like butterflies. Don't chase them. Instead, plant flowers.🦋 #GenshinImpact# #原神5ꓸ7# #丝柯克#
0
20
8.7K
526
Forward to community
The Ecological Restoration and Desertification Control Project in the Aximan Area, Xinjiang, China, serves as a vital ecological barrier protecting the Aksu oasis. Following the same spirit of the Kokyar Sand Control Project, it was officially launched in 2021. People of all ethnic groups have worked side by side. They innovatively use reclaimed urban wastewater instead of precious river water for tree planting, laying dozens of kilometres of water pipelines and drip irrigation networks. Native drought-resistant species such as sacsaoul, tamarisk and sea buckthorn are selected. By combining trees, shrubs and grasses, sand enclosure and afforestation go hand in hand with wetland restoration, building layered shelterbelts to hold back shifting sand. After five years of consistent efforts, large stretches of sand dunes have been stabilized by vegetation, and forest coverage has risen markedly. Deserts have been transformed into wetlands, reed marshes and forest belts. Birds including egrets, wild ducks and black storks have returned to the lake. This green Great Wall against sandstorms across the desert is a vivid example of harmony between humans and nature, and of people from all ethnic groups jointly building a beautiful homeland. 中国新疆艾西曼区域生态修复及荒漠化治理工程是守护阿克苏绿洲的一道关键生态屏障,也是和柯柯牙防风治沙工程一脉相承的荒漠治理工程,于2021年正式启动 。各族群众相守相伴,创新利用城市再生水,铺设数十公里输水管道、滴灌管网,不用珍贵的河水来种树。选用梭梭、红柳、沙棘等本土耐旱植物,乔、灌、草搭配,封沙育林、湿地修复同步推进,筑起层层防风固沙林带,阻挡流沙前移 。 五年持续治理,大片沙丘慢慢被植被固定,森林覆盖率大幅提升。荒漠变成湿地、苇丛与林带,白鹭、野鸭、黑鹳等鸟类重回湖区。大漠之上,这一道防风挡沙的绿色长城,是人与自然共生、各族群众共建美好家园的生动样本。 #大美新疆# #石榴花开映天山# #China# #XinjiangThe#
Show more