README.md

---
license: mit
datasets:
- raincandy-u/TinyChat
language:
- en
tags:
- lm
- 12m
- VDRONT
---

**FP16**


**Ru:**

**VDrontV2-mini** имеет около 10 миллионов параметров и является языковой моделью (LM), обученной на датасете **raincandy-u/TinyChat** до значения потерь (loss) 1.0. Умеет продолжать простые диалоги, подобные тем, что в датасете, но не очень хорошо.

**Пример диалогов (temp 0.7):**
```text
👤 You: <A>Do you like dogs?  <end>\n<B>
🤖 Assistant: Yes, I do. We should like to me. <end>
<A>Why do we have a leader? <end>
<B>Because they are very good at home. <end>
<A>I want to be a bike too. Can you teach me? <end>
<B>Yes, I can. My mom's new. <end>
<A>What does your favorite food? <end>
<B>It's about a bike. <end>
<A>I want to draw a dog. It's kind of my friend. <end>
<B>I want to see a cat. Do you have a dog. But I like it? <end>
<B>Yes, I want to ride a dog. He is a cat. I want to read you. <end>
<A>
<B>Me too! I like to eat a dog. I want to go to be a pet dog. <end>
<A>How many things too. But I like to read it. <end>
<A>I like cars. <end>My family is a pet? <end>
<B>Let's dog
```
use.py - пример чата с ней на температуре 0.4.

En:

VDrontV2-mini has about 10 million parameters and is a language model (LM) trained on the raincandy-u/TinyChat dataset until reaching a loss value of 1.0. It can continue simple conversations similar to those found in the dataset, but not very well.

Example dialogues (temp 0.7):
```text
👤 You: <A>Do you like dogs?  <end>\n<B>
🤖 Assistant: Yes, I do. We should like to me. <end>
<A>Why do we have a leader? <end>
<B>Because they are very good at home. <end>
<A>I want to be a bike too. Can you teach me? <end>
<B>Yes, I can. My mom's new. <end>
<A>What does your favorite food? <end>
<B>It's about a bike. <end>
<A>I want to draw a dog. It's kind of my friend. <end>
<B>I want to see a cat. Do you have a dog. But I like it? <end>
<B>Yes, I want to ride a dog. He is a cat. I want to read you. <end>
<A>
<B>Me too! I like to eat a dog. I want to go to be a pet dog. <end>
<A>How many things too. But I like to read it. <end>
<A>I like cars. <end>My family is a pet? <end>
<B>Let's dog
```
use.py — an example chat session with it at a temperature of 0.4.