wop/Cosmos-SFT
1.57 GB
Wop
wop
AI & ML interests
AI research
AGI
Recent Activity
liked a Space about 4 hours ago
wop/inference updated a Space about 4 hours ago
wop/inference updated a collection about 4 hours ago
EmotionsOrganizations
Posts 1
Post
193
π __Monostep v1__ is up β[Monostep-v1 Demo}( wop/Cosmos-T2-Chat)
A tiny (~16.6M) experimental model that predicts 4 tokens per forward pass instead of one. A Transformer trunk pools the prompt into a single vector, then 4 sequential "slot" heads emit a block of tokens left-to-right β a lightweight take on multi-token prediction.
Trained on GSM8K (GPT-2 tokenizer, 10 epochs). It's small and rough β answers are often wrong β but it's a fun little testbed for block decoding. Weights, config, training curves, and a self-contained inference snippet are all in the repo.
Also wired into the Cosmos T2-Accelerate chat demo, where it streams those 4-token blocks live. π§ͺ
#multitokenprediction #gsm8k #smallmodels
A tiny (~16.6M) experimental model that predicts 4 tokens per forward pass instead of one. A Transformer trunk pools the prompt into a single vector, then 4 sequential "slot" heads emit a block of tokens left-to-right β a lightweight take on multi-token prediction.
Trained on GSM8K (GPT-2 tokenizer, 10 epochs). It's small and rough β answers are often wrong β but it's a fun little testbed for block decoding. Weights, config, training curves, and a self-contained inference snippet are all in the repo.
Also wired into the Cosmos T2-Accelerate chat demo, where it streams those 4-token blocks live. π§ͺ
#multitokenprediction #gsm8k #smallmodels
Articles 2
Article
3
Does Depth Actually Help Reasoning? A Tiny Experiment on 2Γ T4
spaces 62
pinned
Running
1
Distill yourself
π
Annotate text with emotion intensity and export as dataset
pinned
Running
4
Fontend
π
Chat with an AI assistant in real time
pinned
Sleeping
Agents
2
Bg Remover
π’
Remove background from images
Running
1
Inference
π
Detect and visualize emotions in any text
Running
1
Visualize
π
Explore and visualize emotions flowing through text entries
Running
Agents
1
Server Backend
β‘
Chat with the CosmosβT3 language model for instant responses
models 49
wop/Cosmos-T3-BASE-soon
Updated
wop/Monostep-v1
Text Generation β’ Updated β’ 606 β’ 4
wop/Cosmos-T2A-low
Text Generation β’ Updated
wop/Cosmos-T2-Accelerate-Beta2
Text Generation β’ Updated
wop/Cosmos-T2-Accelerate-beta
Text Generation β’ Updated
wop/Cosmos-T2-Accelerate-Preview
Text Generation β’ Updated
wop/Cosmos-T2-80M-Test
Text Generation β’ Updated
wop/Cosmos-T-80M
Text Generation β’ Updated β’ 4
wop/Creativity-lfm2-5-350M
0.4B β’ Updated β’ 30 β’ 2
wop/Opus4Qwen4
4B β’ Updated β’ 37 β’ 2
datasets 18
wop/Human-Essence-Dataset
Viewer β’ Updated β’ 27 β’ 1
wop/debug-self
Viewer β’ Updated β’ 1
wop/smallestdata-sft
Viewer β’ Updated β’ 19 β’ 52 β’ 1
wop/Super-Intelligence-Immersive
Viewer β’ Updated β’ 2 β’ 30
wop/Cosmos-T3-06-09-2026
Viewer β’ Updated β’ 295k β’ 59 β’ 1
wop/miniconveserations
Viewer β’ Updated β’ 16 β’ 32
wop/instantShot
Viewer β’ Updated β’ 177 β’ 74 β’ 1
wop/TinyData
Viewer β’ Updated β’ 1.64k β’ 95 β’ 2
wop/Generic-Chat-Prompts-Tier1
Viewer β’ Updated β’ 172 β’ 21
wop/minitron-dataset
Viewer β’ Updated β’ 288k β’ 113 β’ 1