Cerebras REAP
Collection
Sparse MoE models compressed using REAP (Router-weighted Expert Activation Pruning) method • 30 items • Updated • 152
Jayson
j4ys0n
·
AI & ML interests
None yet
Recent Activity
new activity about 1 month ago
GnLOLot/MiniCPM5-1B-Claude-Opus-Fable5-Thinking:ONNX version? upvoted a collection 6 months ago
Cerebras REAPOrganizations
None yet
ONNX version?
1
#1 opened about 1 month ago
by j4ys0n
upvoted a collection 6 months ago
reacted to mike-ravkine's post with 🔥 7 months ago
Post
3063
Happy 2026 everyone!
I've been busy working on some new ranking/position methodologies and excited to start sharing some results.
Plot legends:
- X = truncation rate (low = good)
- ? = confusion rate (low = good)
- blue bars = average completion tokens (low = good)
- black diamonds = CI-banded performance (high = good)
- cluster squares = models inside this group are equivalent
openai/gpt-oss-120b remains the king in all dimensions of interest: truncation rates, completion lengths and performance. If I had but one complaint it's the reason_effort does not seem to actually work - more on this soon.
Second is a 3-way tie in performance between the Qwen3-235B-2507 we all know and love with an unexpected entrant - ByteDance-Seed/Seed-OSS-36B-Instruct
This is a very capable model and it's reasoning effort controls actually works, but you should absolutely not leave it on the default "unlimited" - enable a sensible limit (4k works well for 8k context length).
Third place is another 3-way tie, this one between Seed-OSS-36B (it straddles the CI boundary between 2nd and 3rd place), Qwen/Qwen3-Next-80B-A3B-Instruct (demonstrating that full attention may be overrated after all and gated is the way to go) and the newly released zai-org/GLM-4.7 which offers excellent across the board performance with some of the shortest reasoning traces I've seen so far.
I've been busy working on some new ranking/position methodologies and excited to start sharing some results.
Plot legends:
- X = truncation rate (low = good)
- ? = confusion rate (low = good)
- blue bars = average completion tokens (low = good)
- black diamonds = CI-banded performance (high = good)
- cluster squares = models inside this group are equivalent
openai/gpt-oss-120b remains the king in all dimensions of interest: truncation rates, completion lengths and performance. If I had but one complaint it's the reason_effort does not seem to actually work - more on this soon.
Second is a 3-way tie in performance between the Qwen3-235B-2507 we all know and love with an unexpected entrant - ByteDance-Seed/Seed-OSS-36B-Instruct
This is a very capable model and it's reasoning effort controls actually works, but you should absolutely not leave it on the default "unlimited" - enable a sensible limit (4k works well for 8k context length).
Third place is another 3-way tie, this one between Seed-OSS-36B (it straddles the CI boundary between 2nd and 3rd place), Qwen/Qwen3-Next-80B-A3B-Instruct (demonstrating that full attention may be overrated after all and gated is the way to go) and the newly released zai-org/GLM-4.7 which offers excellent across the board performance with some of the shortest reasoning traces I've seen so far.
- 1 reply
vLLM MoE Config
#32 opened 9 months ago
by j4ys0n
vLLM MoE config for RTX 5090
#5 opened 10 months ago
by j4ys0n
vLLM MoE config
#5 opened 10 months ago
by j4ys0n
sad! no tool calls in streaming mode.
#15 opened 12 months ago
by j4ys0n
reacted to prithivMLmods's post with ❤️ about 1 year ago
Post
2082
Dropping some image classification models for content moderation and classifiers trained with datasets available on the Hub. All are fine-tuned on the
🤗Models :
AI or Not : prithivMLmods/AIorNot-SigLIP2
Driver Drowsiness Detection : prithivMLmods/DOZE-GUARD-RLDD
Subset 10 ImageNet : prithivMLmods/IMAGENETTE
🥊Datasets :
+ competitions/aiornot
+ akahana/Driver-Drowsiness-Dataset
+ frgfm/imagenette
🔗Collection :
[The previous collection of models is also listed in the same collection, so you can find more models focused on image classification tasks.]
- prithivMLmods/multiclass-image-classification-05142025-68234c8010a9350a4d6739b5
Find a collections inside the collection.🤪👆
To know more about it, visit the model card of the respective model.
siglip2 backbone, (competitions AIOrNot, Imagenette, and Driver-Drowsiness). Models and datasets are listed below:🤗Models :
AI or Not : prithivMLmods/AIorNot-SigLIP2
Driver Drowsiness Detection : prithivMLmods/DOZE-GUARD-RLDD
Subset 10 ImageNet : prithivMLmods/IMAGENETTE
🥊Datasets :
+ competitions/aiornot
+ akahana/Driver-Drowsiness-Dataset
+ frgfm/imagenette
🔗Collection :
[The previous collection of models is also listed in the same collection, so you can find more models focused on image classification tasks.]
- prithivMLmods/multiclass-image-classification-05142025-68234c8010a9350a4d6739b5
Find a collections inside the collection.🤪👆
To know more about it, visit the model card of the respective model.