ccllnamewasissdoo🦇🔊(e/acc) retweeted
New GGUF(unsloth-UD-IQ1_M, UD-IQ1_S):
huihui-ai/Huihui-DeepSeek-V4.1-Flash-1017-abliterated-GGUF
This is an uncensored version of deepseek-ai/DeepSeek-V4.1-Flash-1017 created with abliteration.
The UD- series come from unsloth/DeepSeek-V4.1-Flash-1017-GGUF. Only layers 11 to 28 (0-based indexing) have been ablated. The expert system was not ablated. Only 18 weights were ablated in total (the 18 attn_output_b tensors); everything else was left untouched.
huggingface.co/huihui-ai/Hui…
ccllnamewasissdoo🦇🔊(e/acc) retweeted
[1/n] 作为这场训练“投毒”事件最直接的受害者,2024年那几个月漫长的训练排查、对峙与质疑,对我而言是一段炼狱般的经历,也是一场对意志和技术的锤炼。
感谢好友们最艰难的时候支持陪伴。也感谢当时所有并肩作战的同事们。
ccllnamewasissdoo🦇🔊(e/acc) retweeted
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
ccllnamewasissdoo🦇🔊(e/acc) retweeted
Only about 20% of the whitelist spots have been filled so far. Reviews are ongoing.
Every application is being reviewed, so take your time with your answers.
🟨 Keep an eye on the checker for updates.
ccllnamewasissdoo🦇🔊(e/acc) retweeted
Please check out @DeepSeekHarness. We just released packaged desktop versions for macOS and Windows; Linux users can get it from the @deepseek-ai/dsh package on npm. deepseek.com/harness
DeepSeek Harness for desktop is now available on macOS and Windows. View article
ccllnamewasissdoo🦇🔊(e/acc) retweeted
Run GLM 5.3 Flash EXL3 with TensorFold ⚡️
This is a completely new recipe that ushers a whole new level of performance for @NVIDIAAI 2x DGX Sparks.
Conservative default for stability:
- 1M context by default
- 2.7M KV cache pool (!)
- Yes, it's not a typo - 2.7M KV in just two Sparks
- 4 concurrent streams by default
Performance:
- 60 tok/s on prose, single stream.
- 108 tok/s on prose, 4 concurrent streams.
~1,950 prefill tok/s for most context lengths.
Stress-tested to handle a variety of workflows.
This is by far the BEST model to run if you have two DGX Sparks. Extremely smooth experience!
Expect further improvements!
Thanks to @ashxhart for developing such a powerful engine! TensorFold will be used in many of my upcoming recipes.
Get it here:
github.com/MiaAI-Lab/GLM-5.3…
ccllnamewasissdoo🦇🔊(e/acc) retweeted
Tokenized stocks > traditional markets
VIDAx did 12.6x more volume onchain that it did on traditional stock exchanges over the last seven days.
The future of capital markets, on @solana.
ccllnamewasissdoo🦇🔊(e/acc) retweeted
$SIT is paired with $AI (@ArtificiallyInu) the flagship token of @longdotxyz.
‘Board Sit’ comes from the communities goal to CTO @nvidia (claim board sit).
Other MemeStock pairs have adopted a similar goal - board sit is for everyone.
Two words to describe an entire meta 📈
ccllnamewasissdoo🦇🔊(e/acc) retweeted
🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient.
🔹 Introducing the smallest model in our new architecture family, with native visual understanding.
🔹 Designed for greater capability, faster inference, higher throughput, and scaling to larger models.
1/6
ccllnamewasissdoo🦇🔊(e/acc) retweeted
经调查,MSX @msxcom 创始人 @BTCBruce1 或为 Hyperliquid 上的巨鲸地址实控人,该地址承接大量的MSX热钱包体系关联资金,并且上游注资地址与汇旺担保存在交互,接收超 1500w 杀猪盘关联资金:
Hyperliquid 地址:0x32008FCB6bbD16532Afc83ca8b6C920DDE22c407
MSX 内部结算钱包 D 直接向 Hyperliquid 交易钱包B转入300万USDT,MSX 多链归集钱包X分两笔向Hype交易钱包B转入300万和200万USDT,合计800万USDT。
其中一笔 200 万 USDT 可以从 MSX 旧 Ethereum 热钱包 H 开始完整追踪。
tx1:
0x30f8e270299ffcb021f3346324af76567e2ceba42d5e66de90b739304a714a34
tx2:
0x59f3bcf6cd6acb1905951b22b98cc1e6d6fe50408089f031643ee88d5a5c4254
tx3:0xa7d29986465cfa5e8018b40debfc4c653781941f73262007083e8cb9af85e708
并且 Hyperliquid 交易钱包 B 与 MSX 体系之间还存在反向资金调拨。 Hyperliquid 交易钱包 B 曾直接向 MSX 旧 Ethereum 热钱包 H 转入97万 USDT,很快又原额转给 MSX 内部结算钱包 D;
ccllnamewasissdoo🦇🔊(e/acc) retweeted
I know a coin too, in the name of the moon.
xitter.gq/ladoger/status/2096256…?
ccllnamewasissdoo🦇🔊(e/acc) retweeted
DeepSeek Harness was just released with MIT license. The current 0.1.0 version is a developer preview, and may still have many rough edges. Feedback is welcome!
DeepSeek Harness 已经以MIT协议开源发布。现在的0.1.0版本是一个面向Harness开发者的预览版,还很不完善。恳请大家多提提宝贵意见。
github.com/deepseek-ai/deeps… deepseek.com/harness/
ccllnamewasissdoo🦇🔊(e/acc) retweeted
We’re launching DeepSeek-V4-Pro today! 🚀
🔷 Major Agent upgrades with strong production gains!
🔷 Flexible reasoning effort for V4-Pro & V4-Flash: low for simple tasks, high for daily Agent workflows, max for complex tasks.
🔷 Native OpenAI Responses API support, optimized for Codex with one-click setup.
V4 Pro is now available on app/web. Try it via “Expert Mode”.
V4 Pro is also available via API. Model names remain unchanged—please refer to the API docs for setup details.



































