Mayank Maheshwari

mayankiit04's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

New activity in unsloth/Qwen3.8-Flash-Next-GGUF about 9 hours ago

unsloth studio vs strata engine for qwen3.8-flash-next

2

#80 opened 1 day ago by

mayankiit04

New activity in unsloth/Qwen3.8-Flash-Next-GGUF 1 day ago

llama.cpp now supports MTP in main, but these quants don't seem compatible

2

#79 opened 1 day ago by

FlorinAndrei

New activity in unsloth/Qwen3.8-Flash-Next-GGUF 2 days ago

will the new llama.cpp build 11330+ support the mtp file?

🔥👍 2

3

#78 opened 2 days ago by

mayankiit04

New activity in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF 2 days ago

unsloth shared MTP not working with this IQ3_S on llama.cpp with build 11330

👍 2

1

#36 opened 2 days ago by

mayankiit04

New activity in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF 4 days ago

Benchmark & Reproduction Setup: Qwen3.8-Flash-Next GSQ-RCO IQ3_XXS (24.82 t/s at full 262k context on single RTX 3090)

6

#33 opened 4 days ago by

AMonrroy

New activity in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF 7 days ago

IQ3_S

🔥 6

17

#26 opened 8 days ago by

DzmitryTheOtherOne

New activity in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-Coder-GGUF 7 days ago

What is the SWE Verfied of the the pure GSQ RCO non coder Q3_S

1

#2 opened 7 days ago by

mayankiit04

New activity in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF 14 days ago

Quick Comparison: ISTA-DASLab GSQ-RCO-IQ3_XXS vs unsloth UD-IQ4_XS on 4070Ti Super 16G + 64G RAM

15

#18 opened 16 days ago by

BipedalBit

New activity in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF 18 days ago

what should be the parametric setting to get the model tunned for STEM [GPQA-Diamond/LCB/AIME]

1

#8 opened 18 days ago by

mayankiit04

liked a model 27 days ago

ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF

Image-Text-to-Text • 27B • Updated Sep 2 • 1.64M • 1.95k

New activity in unsloth/Qwen3.8-Flash-Next-GGUF about 1 month ago

Quick share: Got 30–40 t/s on the RTX 5060 Ti 16GB

12

#61 opened about 1 month ago by

QilinWan

Small gain with MTP

7

#59 opened about 1 month ago by

OmarColocci

can we have a gguf varity where ngram layer 2 is at q8?

#53 opened about 1 month ago by

mayankiit04

How can I improve the prefilling speed for this model?

14

#49 opened about 1 month ago by

BipedalBit

New activity in tencent/Hy4-preview about 1 month ago

With Qwen and GLM now in sub 300B local machine level, this is a big model.

👍 1

1

#10 opened about 1 month ago by

mayankiit04

New activity in unsloth/Qwen3.8-Flash-Next-GGUF about 1 month ago

Is the n-gram chunk embedded in the gguf(s), and is it ~51GB independent of quantization?

14

#35 opened about 1 month ago by

dagb

How to keep n-gram table on fast nvme ssd

24

#23 opened about 1 month ago by

mayankiit04

New activity in Qwen/Qwen3.8-Flash-Next about 1 month ago

Wow this is an outright killer model... anyone can now run this with mere 48 gb ram/vram & a fast nvme drive

6

#30 opened about 1 month ago by

mayankiit04

here qwen3.8-flash-next is 125B but unsloth has 180B ... how come?

3

#17 opened about 1 month ago by

mayankiit04

New activity in ornith-ai/Ornith-1.5-397B about 1 month ago

Are you serious, Chatgpt puts this model at estimate 59 score AA above opus 4.8!!

👍 1

#3 opened about 1 month ago by

mayankiit04