Niklas
@niklaslenz_ai • 2,288 subscribers
I run local AI at home on 4 DGX Sparks, 1 M5U 256GB, 1 RTX 5090, 2 RTX 5060 Ti 16GB, and rank what runs best on the machine you own. LLMs, image/video gen. etc
Videos
🚨 This is huge. Fish Audio just opened the weights of Drama 3, and it's the first open TTS model that sounds close to perfect to me. Other models can switch emotion mid-sentence. Drama 3 is the first one I've heard where it actually sounds right. Three tests against Qwen3-TTS, each on one DGX Spark at home: happy turning sad mid-sentence, a sad whisper, and pure excitement. Listen to where the mood drops. I think this kills ElevenLabs for anyone who has their own compute. The catch: it's a 35B model (3B active), a 41.5 GB download, and on the Spark it only just keeps up with real time: 10 s of audio take 9.7 s (Qwen3-TTS 7.6 s).
Niklas150,819 次观看 • 16 天前
Incredible! I have never seen a better music machine for my prompt, Opus and Sonnet included. 9/10. Turn your sound on. DeepSeek-V4.1-Flash-1017 built it on four DGX Sparks in 93.9 min, with Mia's TensorFold recipe. I checked it myself: this is not an Opus output. It ran on my own Sparks. Setup: – Recipe: Mia's, commit 6beaad9, with the 8 TF_DS_* switches her README table sets on; one model split across all four Sparks (TP4), 16 streams – Context: 1,048,576 tokens – Agent: DeepSeek Harness 0.2.1-rc.2, one prompt, one try, no help, thinking on – Sampling from the DeepSeek model card: temperature 1.0, top_p 1.0, top_k 20 – The sandbox had no internet; all 71 requests went to my local server Speed: – Decode, code: 69.5 tok/s (separate speed run, thinking off) – Decode, prose: 70.2 tok/s – 500 words: about 10.0 s – Eight requests at once: 139.0 tok/s in total The machine: – Built in 93.9 min, 272,026 tokens – Browser check: 9.0/10 – 18 voices, a wheel with 720 sockets, 6 marble types, 4 pieces – All sound is synthesized live in the browser, no samples – Sound on: the music in the video is what the page plays itself Good to know: – After 90 minutes the harness told the model to finish, and it was done at 93.9. – The browser check took 1 point off for the frame-rate test (under 3 fps, it renders in software). – The recipe's default answer budget is 32,768 tokens, the model card allows 262,144, so I raised the limit to that. – On her own README method my ring gets 119.9 tok/s on code, hers is 117.4. Her README lists 165.3 tok/s at eight requests (prose, thinking on, temperature 0.7); my speed test runs thinking off at the model card's sampling. – The ranking card at the end is the class ranking from my voxel-world test. Play the music machine in your browser, every number and the raw output: Mia's recipe and the exact setup: Recipe by Mia. TensorFold engine by Ash Hart. Model by DeepSeek.
Niklas23,772 次观看 • 3 天前
GLM-5.3-Flash on two DGX Sparks built this voxel sky kingdom in 71.1 min, on Mia's recipe and the TensorFold Python engine. Mia's recipe at commit 3fd60dfc: EXL3 4 bpw quant, one model split across both Sparks (TP2), DFlash2 draft model, 8 slots (her recipe allows up to 8; the default on two Sparks is 4). Agent: OpenCode 1.19.3. One prompt, one try, no help. Thinking on, sampling from the model card: temperature 1.0, top_p 0.95. Voxel world: done in 71.1 min Tokens written: 109,305 Browser check: 9/10 Decode, code: 51.2 tok/s (separate speed run, thinking off) Decode, prose: 48.6 tok/s 500 words: about 14.0 s Eight requests at once: 99.2 tok/s in total Context: 1,048,576 tokens, the recipe's setting Aetherfall: floating isles over a cloud sea with a sky whale and its calf, a waterfall, a windmill, a castle, a shrine, airships and its own camera tour. Open: parts of the world are cut off hard at the edges, most likely a bug in the generated code. Answer limit in the build test: 65,536 tokens, the figure names for its Terminal-Bench runs. Its model card names up to 163,840 for another test. The longest answer was 54,782 tokens, so none was cut. Before this I ran the same task on Mia's Zig-engine branch, which her docs call work in progress: Decode, code: 26.6 tok/s Decode, prose: 25.5 tok/s Voxel world: done in 98.8 min Browser check: 8.0/10 The Python engine is the one her main recipe runs. Play the world in your browser. Every number, the exact command and the raw output: Recipe and quant by Mia. TensorFold engine by Ash Hart. GLM-5.3-Flash by Z.ai. DFlash2 draft by Inco AI.
Niklas16,730 次观看 • 6 天前
DeepSeek-V4.1-Flash-1017 decodes code at 80.5 tok/s on one Mac Studio M5 Ultra 256 GB with BobClawblaw's recipe on TensorFold. Setup: – Recipe and TensorFold branch deepseek-v41-main (72a552ed), his own tool and settings – Context 524,288 tokens, 4 streams, DSpark drafts on, temperature 0 – Speed and quality with thinking off, except GPQA with thinking Speed: – Decode, code: 80.5 tok/s – Decode, prose: 70.5 tok/s – 500 words: about 9.3 s – First token after a 32K prompt: 18.6 s – 8 requests at once: 88.3 tok/s in total Quality: – MMLU-Pro: 67.5% ± 4.6 points – GPQA Diamond: 45.0% ± 6.9 points – IFEval: 90.5% ± 4.1 points – LiveCodeBench: 60.0% ± 14.5 points – GPQA with thinking: 76.7% ± 14.6 points Voxel world: – Built as an agent in 45.7 min, DeepSeek Harness 0.2.1-rc.2, one prompt, one try, no help, thinking on – Tokens written: 112,065 – Browser check: 8.5/10 Good to know: – The author runs temperature 0, the DeepSeek card suggests 1.0. I kept the author's setting. – The browser check ran on a Linux box from the same files. – Answers cut at the cap count as wrong: LiveCodeBench 14 of 40, GPQA Diamond 9 of 198, MMLU-Pro 1 of 400. Play the world in your browser. All numbers, the exact command and the raw output: Recipe by BobClawblaw on Ash Hart's TensorFold, model by DeepSeek.
Niklas11,028 次观看 • 4 天前
DeepSeek-V4.1-Flash-1017 on 4 DGX Sparks vs 2 DGX Sparks, same voxel-world prompt: the 2-Spark run finished in 59.7 minutes. Both recipes are by Mia. 4 Sparks runs her TP4 recipe on SGLang. 2 Sparks runs her recipe on Ash Hart's TensorFold. Not just a box count: 4 Sparks: official checkpoint, sampling from the model card. 2 Sparks: EXL3 2.9 bpw quant, sampling from Mia's recipe. The answer limit was raised to 262,144 tokens on both. 4 Sparks, world Lumenfall: Agent time: 90.8 min Tokens written: 196,854 Browser check: 9/10 Decode (speed test): 51 tok/s 2 Sparks, world Aetherfall: Agent time: 59.7 min Tokens written: 131,268 Browser check: 8/10 Decode (speed test): 67 tok/s Context tested on both: 128K. Open: only the 4-Spark run reached the 90 minute mark and got the automatic finish-now note. Open: 4-Spark decode on the code prompt swung between two runs (36 and 65 tok/s), cause not found. The 2-Spark world is almost as good. Really good. Play both worlds in your browser: 4 Sparks: 2 Sparks:
Niklas15,876 次观看 • 9 天前
没有更多内容可加载