Update README.md
Browse files
README.md
CHANGED
@@ -9,26 +9,22 @@ language:
|
|
9 |
- en
|
10 |
---
|
11 |
|
12 |
-
#
|
13 |
|
14 |
## Overview
|
15 |
|
16 |
-
LLama3-Gaja-Hindi-8B-v0.1 is an extension of the Ambari series, a bilingual English/Hindi model developed and released by [Cognitivelab.in](https://www.cognitivelab.in/). This model is specialized for natural language understanding tasks, particularly in the context of instructional pairs. It is built upon the [Llama3 8b](https://huggingface.co/meta-llama/Meta-Llama-3-8B) model, utilizing a fine-tuning process with a curated dataset of translated instructional pairs.
|
17 |
-
|
18 |
-
<img src="https://cdn-uploads.huggingface.co/production/uploads/6442d975ad54813badc1ddf7/G0u9L6RQJFinST0chQmfL.jpeg" width="500px">
|
19 |
-
|
20 |
## Generate
|
21 |
```python
|
22 |
import torch
|
23 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
24 |
from transformers import GenerationConfig, TextStreamer , TextIteratorStreamer
|
25 |
|
26 |
-
model = AutoModelForCausalLM.from_pretrained("
|
27 |
-
tokenizer = AutoTokenizer.from_pretrained("
|
28 |
|
29 |
# Existing messages list
|
30 |
messages = [
|
31 |
-
{"role": "system", "content": " You are
|
32 |
{"role": "user", "content": "Who are you"}
|
33 |
]
|
34 |
|
@@ -54,19 +50,19 @@ print(tokenizer.decode(response, skip_special_tokens=True))
|
|
54 |
|
55 |
## Multi-turn Chat
|
56 |
|
57 |
-
To use the
|
58 |
|
59 |
```python
|
60 |
import torch
|
61 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
62 |
from transformers import GenerationConfig, TextStreamer , TextIteratorStreamer
|
63 |
|
64 |
-
model = AutoModelForCausalLM.from_pretrained("
|
65 |
-
tokenizer = AutoTokenizer.from_pretrained("
|
66 |
|
67 |
# Existing messages list
|
68 |
messages = [
|
69 |
-
{"role": "system", "content": " You are
|
70 |
]
|
71 |
|
72 |
# Function to add user input and generate response
|
@@ -147,7 +143,7 @@ while True:
|
|
147 |
|
148 |
## Prompt formate
|
149 |
|
150 |
-
system prompt = `You are
|
151 |
```
|
152 |
<|begin_of_text|><|start_header_id|>system<|end_header_id|>
|
153 |
|
@@ -162,11 +158,3 @@ system prompt = `You are Gaja, an AI assistant created by Cognitivelab and train
|
|
162 |
|
163 |
## Benchmarks
|
164 |
coming soon
|
165 |
-
|
166 |
-
## Bilingual Instruct Fine-tuning
|
167 |
-
|
168 |
-
The model underwent a pivotal stage of supervised fine-tuning with low-rank adaptation, focusing on bilingual instruct fine-tuning. This approach involved training the model to respond adeptly in either English or Hindi based on the language specified in the user prompt or instruction.
|
169 |
-
|
170 |
-
## References
|
171 |
-
|
172 |
-
- [Ambari-7B-Instruct Model](https://huggingface.co/Cognitive-Lab/Ambari-7B-Instruct-v0.1)
|
|
|
9 |
- en
|
10 |
---
|
11 |
|
12 |
+
# Eli
|
13 |
|
14 |
## Overview
|
15 |
|
|
|
|
|
|
|
|
|
16 |
## Generate
|
17 |
```python
|
18 |
import torch
|
19 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
20 |
from transformers import GenerationConfig, TextStreamer , TextIteratorStreamer
|
21 |
|
22 |
+
model = AutoModelForCausalLM.from_pretrained("Neohumans-ai/Eli", torch_dtype=torch.bfloat16).to("cuda")
|
23 |
+
tokenizer = AutoTokenizer.from_pretrained("Neohumans-ai/Eli", trust_remote_code=True)
|
24 |
|
25 |
# Existing messages list
|
26 |
messages = [
|
27 |
+
{"role": "system", "content": " You are Eli, an AI assistant created by NeoHumans-ai and trained on top of Llama 3 Large language model (LLM), proficient in English and Hindi. You can respond in both languages based on the user's request."},
|
28 |
{"role": "user", "content": "Who are you"}
|
29 |
]
|
30 |
|
|
|
50 |
|
51 |
## Multi-turn Chat
|
52 |
|
53 |
+
To use the Eli model, you can follow the example code below:
|
54 |
|
55 |
```python
|
56 |
import torch
|
57 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
58 |
from transformers import GenerationConfig, TextStreamer , TextIteratorStreamer
|
59 |
|
60 |
+
model = AutoModelForCausalLM.from_pretrained("Neohumans-ai/Eli", torch_dtype=torch.bfloat16).to("cuda")
|
61 |
+
tokenizer = AutoTokenizer.from_pretrained("Neohumans-ai/Eli", trust_remote_code=True)
|
62 |
|
63 |
# Existing messages list
|
64 |
messages = [
|
65 |
+
{"role": "system", "content": " You are Eli, an AI assistant created by NeoHumans-ai and trained on top of Llama 3 Large language model (LLM), proficient in English and Hindi. You can respond in both languages based on the user's request."},
|
66 |
]
|
67 |
|
68 |
# Function to add user input and generate response
|
|
|
143 |
|
144 |
## Prompt formate
|
145 |
|
146 |
+
system prompt = `You are Eli, an AI assistant created by NeoHumans-ai and trained on top of Llama 3 Large language model(LLM), proficient in English and Hindi. You can respond in both languages based on the users request.`
|
147 |
```
|
148 |
<|begin_of_text|><|start_header_id|>system<|end_header_id|>
|
149 |
|
|
|
158 |
|
159 |
## Benchmarks
|
160 |
coming soon
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|