Hugging Face's logo

dotlabs
/
void.1

Text Generation
Transformers
TensorBoard
Safetensors
English
qwen3
byte-level
pretraining
symbolic
text-generation-inference
Model card Files
xet
 Metrics Community
3
appvoid commited on
Commit
9015a72
·
verified ·
1 Parent(s): 6f41f7a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +8 -10
README.md CHANGED
@@ -8,17 +8,15 @@ tags: [qwen3, byte-level, pretraining]
8
 
9
  State of the art, small language model pretrained from scratch on a diverse set of high-quality text and our internal symbolic kernel.
10
 
11
- ## Results
12
 
13
- | Benchmark | Score |
14
- | ---------------------------- | ------- |
15
- | HellaSwag | 38.68% |
16
- | PIQA | 67.46% |
17
- | ARC-Easy | 47.31% |
18
- | ARC-Challenge | 28.16% |
19
- | ArithMark-3 | 44.80% |
20
- | Intelligence Index | 23.9182 |
21
- | Benchmark Max Length | 2,048 |
22
 
23
 
24
  We used the revision on step 900,000 for evaluations which trained for around 120 billion bytes which is around 30 to 35 billion bpe tokens. For more details on the evals and inference, please have a look at the [official notebook](https://huggingface.co/dotlabs/void.1/blob/main/evals.ipynb).
 
8
 
9
  State of the art, small language model pretrained from scratch on a diverse set of high-quality text and our internal symbolic kernel.
10
 
11
+ ## Comparison
12
 
13
+ | Model | Params | HellaSwag | PIQA | ARC-Easy | ARC-Challenge | ArithMark-3 | Intelligence Index |
14
+ | ----------------------------------------------------------------- | ------ | --------- | ------ | -------- | ------------- | ----------- | ------------------ |
15
+ | [void.1](https://huggingface.co/dotlabs/void.1)\* | **90.15M** | **38.68%** | **67.46%** | 47.31% | **28.16%** | **44.80%** | **23.92** |
16
+ | [100M-exp](https://huggingface.co/User01110/100M-exp) | 98.16M | 37.78% | 66.97% | **49.83%** | 27.22% | 40.00% | 22.47 |
17
+ | [Rose-1.5-Medium](https://huggingface.co/GODELEV/Rose-1.5-Medium) | 98.28M | 38.09% | 64.80% | 47.22% | 27.13% | 40.70% | 21.07 |
18
+ | [tinctura-v1](https://huggingface.co/bench-labs/tinctura-v1) | 96.2M | 37.96% | 65.61% | 47.98% | 25.77% | 38.40% | 20.81 |
19
+ | [Surjo-100m](https://huggingface.co/SurjoLabs/Surjo-100m) | 97.7M | 35.05% | 63.87% | 47.64% | 25.85% | 38.90% | 18.86 |
 
 
20
 
21
 
22
  We used the revision on step 900,000 for evaluations which trained for around 120 billion bytes which is around 30 to 35 billion bpe tokens. For more details on the evals and inference, please have a look at the [official notebook](https://huggingface.co/dotlabs/void.1/blob/main/evals.ipynb).