Text Classification
Transformers
Safetensors
MLX
English
qwen3
feature-extraction
tinyjev
jev
decision-model
system-one
typed-decisions
text-embeddings-inference
Instructions to use AnkitAI/TinyJev-0.6B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AnkitAI/TinyJev-0.6B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="AnkitAI/TinyJev-0.6B")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("AnkitAI/TinyJev-0.6B") model = AutoModel.from_pretrained("AnkitAI/TinyJev-0.6B", device_map="auto") - MLX
How to use AnkitAI/TinyJev-0.6B with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download AnkitAI/TinyJev-0.6B --local-dir TinyJev-0.6B
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
demos: batch race and single-question race against GPT-6 Sol
Browse files
README.md
CHANGED
|
@@ -49,12 +49,34 @@ else, because it never generates text; it scores the options you gave it and sto
|
|
| 49 |
|
| 50 |
## Watch it decide
|
| 51 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
<div align="center">
|
| 53 |
<img alt="TinyJev triaging support tickets" src="https://raw.githubusercontent.com/ankit-aglawe/tinyjev/main/assets/demo_triage.gif" width="860">
|
| 54 |
</div>
|
| 55 |
|
| 56 |
Eight real support tickets, three questions each in a single forward pass, about 110 ms per ticket on a base M1. Every number in that recording came from a live run.
|
| 57 |
|
|
|
|
|
|
|
| 58 |
<details>
|
| 59 |
<summary><b>And, for fun, Doom</b> — click to expand</summary>
|
| 60 |
|
|
|
|
| 49 |
|
| 50 |
## Watch it decide
|
| 51 |
|
| 52 |
+
<div align="center">
|
| 53 |
+
<img alt="TinyJev answering 12 typed decisions about one support ticket in one forward pass, next to GPT-6 Sol writing the same answers as JSON token by token" src="https://raw.githubusercontent.com/ankit-aglawe/tinyjev/main/assets/demo_batch.gif" width="860">
|
| 54 |
+
</div>
|
| 55 |
+
|
| 56 |
+
One support ticket, twelve typed decisions: which team, what priority, what the customer wants. TinyJev scores every option of every question in a single forward pass and returns all twelve together, 596 ms on a base M1, 10 of 12 right. GPT-6 Sol writes the same twelve as JSON, token by token: 2,198 ms, 12 of 12 right. Both lanes are real runs replayed at real speed; the ticket, the questions, the expected answers and both measurements are in the [repository](https://github.com/ankit-aglawe/tinyjev).
|
| 57 |
+
|
| 58 |
+
<details>
|
| 59 |
+
<summary><b>Six single questions against GPT-6 Sol</b> — click to expand</summary>
|
| 60 |
+
|
| 61 |
+
<div align="center">
|
| 62 |
+
<img alt="TinyJev and GPT-6 Sol answering the same six never-seen questions from a shared start" src="https://raw.githubusercontent.com/ankit-aglawe/tinyjev/main/assets/demo_race.gif" width="860">
|
| 63 |
+
</div>
|
| 64 |
+
|
| 65 |
+
Six decisions from six domains it never trained on, the same question to both at the same instant. TinyJev 6 of 6 at 86 ms a question; GPT-6 Sol 6 of 6 at 2,042 ms.
|
| 66 |
+
|
| 67 |
+
</details>
|
| 68 |
+
|
| 69 |
+
<details>
|
| 70 |
+
<summary><b>Eight tickets, three questions each</b> — click to expand</summary>
|
| 71 |
+
|
| 72 |
<div align="center">
|
| 73 |
<img alt="TinyJev triaging support tickets" src="https://raw.githubusercontent.com/ankit-aglawe/tinyjev/main/assets/demo_triage.gif" width="860">
|
| 74 |
</div>
|
| 75 |
|
| 76 |
Eight real support tickets, three questions each in a single forward pass, about 110 ms per ticket on a base M1. Every number in that recording came from a live run.
|
| 77 |
|
| 78 |
+
</details>
|
| 79 |
+
|
| 80 |
<details>
|
| 81 |
<summary><b>And, for fun, Doom</b> — click to expand</summary>
|
| 82 |
|