Skip to content

Commit b3b918d

Browse files
heiruwuheiruwu
authored
docs(readme): populate models' readme (#22)
Because - Introduce each Instill Model compatible implementation of various models This commit - populate models' readme
1 parent b531e68 commit b3b918d

File tree

15 files changed

+1108
-27
lines changed

15 files changed

+1108
-27
lines changed

‎README.md‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -8,6 +8,7 @@ We have a diverse set of models, each optimized for different AI tasks. Please r
88

99
| Model Name | Task Type | Description |
1010
| -------------------------------------------------------------- | --------------------- | ---------------------------------------------------------------------------------------------------------------------- |
11+
| [phi-3.5-vision-instruct](./phi-3-5-vision/README.md) | Chat | Phi-3.5-vision is a lightweight, state-of-the-art open multimodal model. |
1112
| [gte-Qwen2-1.5B-instruct](./gte-Qwen2-1.5B-instruct/README.md) | Embedding | gte-Qwen2-1.5B-instruct is the latest model in the gte (General Text Embedding) model family. |
1213
| [jina-clip-v1](./jina-clip-v1/README.md) | Embedding | jina-clip-v1 is a state-of-the-art English multimodal (text-image) embedding model. |
1314
| [llama2-7b-chat](./llama2-7b-chat/README.md) | Chat | llama2-7b-chat is optimized for dialogue use cases. |

‎gte-Qwen2-1.5B-instruct/README.md‎

Lines changed: 84 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,84 @@
1+
# gte Qwen2 1.5B Instruct
2+
3+
## 📖 Introduction
4+
5+
[gte-Qwen2-1.5B-instruct](https://huggingface.co/Alibaba-NLP/gte-Qwen2-1.5B-instruct) is the latest model in the gte (General Text Embedding) model family. The model is built on Qwen2-1.5B LLM model and use the same training data and strategies as the gte-Qwen2-7B-instruct model.
6+
7+
| Task Type | Description |
8+
| ------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------- |
9+
| [Embedding](https://www.instill.tech/docs/model/ai-task#embedding) | A task to generate means of representing objects like text, images and audio as points in a continuous vector space. |
10+
11+
## 🔄 Compatibility Matrix
12+
13+
To ensure smooth integration, please refer to the compatibility matrix below. It outlines the compatible versions of the model, [`instill-core`](https://github.com/instill-ai/instill-core), and the [`python-sdk`](https://github.com/instill-ai/python-sdk).
14+
15+
| Model Version | Instill-Core Version | Python-SDK Version |
16+
| ------------- | -------------------- | ------------------ |
17+
| v0.1.0 | >v0.39.0-beta | >0.11.0 |
18+
19+
> **Note:** Always ensure that you are using compatible versions to avoid unexpected issues.
20+
21+
## 🚀 Preparation
22+
23+
Follow [this](../README.md) guide to get your custom model up and running! But before you do that, please read through the following sections to have all the necessary files ready.
24+
25+
#### Install Python SDK
26+
27+
Install the compatible [`python-sdk`](https://github.com/instill-ai/python-sdk) version according to the compatibility matrix:
28+
29+
```bash
30+
pip install instill-sdk=={version}
31+
```
32+
33+
#### Get model weights
34+
35+
To download the fine-tuned model weights, please execute the following command:
36+
37+
```bash
38+
git clone https://huggingface.co/Alibaba-NLP/gte-Qwen2-1.5B-instruct
39+
```
40+
41+
## Test model image
42+
43+
After you've built the model image, and before pushing the model onto any **Instill Core** instance, you can test if the model can be successfully run locally first, by running the following command:
44+
45+
```bash
46+
instill run instill-ai/gte-qwen2-1.5b-instruct -g -i '{"prompt": "hi"}'
47+
```
48+
49+
The input payload should strictly follow the the below format
50+
51+
```json
52+
{
53+
"prompt": "..."
54+
}
55+
```
56+
57+
A successful response will return a similar output to that shown below.
58+
59+
```bash
60+
2024-09-11 02:36:18,416.416 INFO [Instill] Starting model image...
61+
2024-09-11 02:36:29,444.444 INFO [Instill] Deploying model...
62+
2024-09-11 02:37:10,118.118 INFO [Instill] Running inference...
63+
2024-09-11 02:37:11,585.585 INFO [Instill] Outputs:
64+
[{'data': {'embeddings': [{'created': 1725993431,
65+
'index': 0,
66+
'vector': [0.014460443519055843,
67+
0.0885428711771965,
68+
0.02166132815182209,
69+
...
70+
...
71+
...
72+
-0.02432815358042717]}]}}]
73+
2024-09-11 02:39:58,651.651 INFO [Instill] Done
74+
```
75+
76+
Here is the list of flags supported by `instill run` command
77+
78+
- -t, --tag: tag for the model image, default to `latest`
79+
- -g, --gpu: to pass through GPU from host into container or not, depends on if `gpu` is enabled in the config.
80+
- -i, --input: input in json format
81+
82+
---
83+
84+
Happy Modeling! 💡

‎jina-clip-v1/README.md‎

Lines changed: 94 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,94 @@
1+
# Jina CLIP V1
2+
3+
## 📖 Introduction
4+
5+
[jina-clip-v1](https://huggingface.co/jinaai/jina-clip-v1) is a state-of-the-art English multimodal (text-image) embedding model.
6+
7+
| Task Type | Description |
8+
| ------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------- |
9+
| [Embedding](https://www.instill.tech/docs/model/ai-task#embedding) | A task to generate means of representing objects like text, images and audio as points in a continuous vector space. |
10+
11+
## 🔄 Compatibility Matrix
12+
13+
To ensure smooth integration, please refer to the compatibility matrix below. It outlines the compatible versions of the model, [`instill-core`](https://github.com/instill-ai/instill-core), and the [`python-sdk`](https://github.com/instill-ai/python-sdk).
14+
15+
| Model Version | Instill-Core Version | Python-SDK Version |
16+
| ------------- | -------------------- | ------------------ |
17+
| v0.1.0 | >v0.39.0-beta | >0.11.0 |
18+
19+
> **Note:** Always ensure that you are using compatible versions to avoid unexpected issues.
20+
21+
## 🚀 Preparation
22+
23+
Follow [this](../README.md) guide to get your custom model up and running! But before you do that, please read through the following sections to have all the necessary files ready.
24+
25+
#### Install Python SDK
26+
27+
Install the compatible [`python-sdk`](https://github.com/instill-ai/python-sdk) version according to the compatibility matrix:
28+
29+
```bash
30+
pip install instill-sdk=={version}
31+
```
32+
33+
#### Get model weights
34+
35+
To download the fine-tuned model weights, please execute the following command:
36+
37+
```bash
38+
git clone https://huggingface.co/jinaai/jina-clip-v1
39+
```
40+
41+
## Test model image
42+
43+
After you've built the model image, and before pushing the model onto any **Instill Core** instance, you can test if the model can be successfully run locally first, by running the following command:
44+
45+
```bash
46+
instill run instill-ai/jina-clip-v1 -g -i '{"text" : "hi", "image": "https://artifacts.instill.tech/imgs/bear.jpg"}'
47+
```
48+
49+
The input payload should strictly follow the the below format
50+
51+
```json
52+
{
53+
"text": "...",
54+
"image": "https://",
55+
}
56+
```
57+
58+
A successful response will return a similar output to that shown below.
59+
60+
```bash
61+
2024-09-11 02:42:38,605.605 INFO [Instill] Starting model image...
62+
2024-09-11 02:42:49,440.440 INFO [Instill] Deploying model...
63+
2024-09-11 02:43:16,851.851 INFO [Instill] Running inference...
64+
2024-09-11 02:43:19,756.756 INFO [Instill] Outputs:
65+
[{'data': {'embeddings': [{'created': 1725993799,
66+
'index': 0,
67+
'vector': [-0.042002271860837936,
68+
0.002093376824632287,
69+
0.007119686808437109,
70+
...
71+
...
72+
...
73+
-0.0350787378847599]},
74+
{'created': 1725993799,
75+
'index': 1,
76+
'vector': [-0.07706715911626816,
77+
-0.006987405009567738,
78+
0.0100631695240736,
79+
...
80+
...
81+
...
82+
-0.02432815358042717]}]}}]
83+
2024-09-11 02:43:23,487.487 INFO [Instill] Done
84+
```
85+
86+
Here is the list of flags supported by `instill run` command
87+
88+
- -t, --tag: tag for the model image, default to `latest`
89+
- -g, --gpu: to pass through GPU from host into container or not, depends on if `gpu` is enabled in the config.
90+
- -i, --input: input in json format
91+
92+
---
93+
94+
Happy Modeling! 💡

‎llama2-7b-chat/README.md‎

Lines changed: 83 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,83 @@
1+
# Llama2 7B Chat
2+
3+
## 📖 Introduction
4+
5+
[Llama 2](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) is a collection of pretrained and fine-tuned generative text models ranging in scale from 7 billion to 70 billion parameters. This is the repository for the 7B fine-tuned model, optimized for dialogue use cases and converted for the Hugging Face Transformers format.
6+
7+
| Task Type | Description |
8+
| -------------------------------------------------------- | ------------------------------------------------------------------------------------------- |
9+
| [Chat](https://www.instill.tech/docs/model/ai-task#chat) | A task to generate conversational style text output base on single or multi-modality input. |
10+
11+
## 🔄 Compatibility Matrix
12+
13+
To ensure smooth integration, please refer to the compatibility matrix below. It outlines the compatible versions of the model, [`instill-core`](https://github.com/instill-ai/instill-core), and the [`python-sdk`](https://github.com/instill-ai/python-sdk).
14+
15+
| Model Version | Instill-Core Version | Python-SDK Version |
16+
| ------------- | -------------------- | ------------------ |
17+
| v0.1.0 | >v0.39.0-beta | >0.11.0 |
18+
19+
> **Note:** Always ensure that you are using compatible versions to avoid unexpected issues.
20+
21+
## 🚀 Preparation
22+
23+
Follow [this](../README.md) guide to get your custom model up and running! But before you do that, please read through the following sections to have all the necessary files ready.
24+
25+
#### Install Python SDK
26+
27+
Install the compatible [`python-sdk`](https://github.com/instill-ai/python-sdk) version according to the compatibility matrix:
28+
29+
```bash
30+
pip install instill-sdk=={version}
31+
```
32+
33+
#### Get model weights
34+
35+
To download the fine-tuned model weights, please execute the following command:
36+
37+
```bash
38+
git clone https://huggingface.co/meta-llama/Llama-2-7b-chat-hf
39+
```
40+
41+
## Test model image
42+
43+
After you've built the model image, and before pushing the model onto any **Instill Core** instance, you can test if the model can be successfully run locally first, by running the following command:
44+
45+
```bash
46+
instill run instill-ai/llama2-7b-chat -g -i '{"prompt": "hi"}'
47+
```
48+
49+
The input payload should strictly follow the the below format
50+
51+
```json
52+
{
53+
"prompt": "..."
54+
}
55+
```
56+
57+
A successful response will return a similar output to that shown below.
58+
59+
```bash
60+
2024-09-11 01:44:12,423.423 INFO [Instill] Starting model image...
61+
2024-09-11 01:44:22,843.843 INFO [Instill] Deploying model...
62+
2024-09-11 01:44:52,935.935 INFO [Instill] Running inference...
63+
2024-09-11 01:44:56,534.534 INFO [Instill] Outputs:
64+
[{'data': {'choices': [{'created': 1725990296,
65+
'finish-reason': 'length',
66+
'index': 0,
67+
'message': {'content': "Hello! It's nice to meet you. "
68+
'Is there something I can help '
69+
'you with or would you like to '
70+
'chat?',
71+
'role': 'assistant'}}]}}]
72+
2024-09-11 01:45:00,240.240 INFO [Instill] Done
73+
```
74+
75+
Here is the list of flags supported by `instill run` command
76+
77+
- -t, --tag: tag for the model image, default to `latest`
78+
- -g, --gpu: to pass through GPU from host into container or not, depends on if `gpu` is enabled in the config.
79+
- -i, --input: input in json format
80+
81+
---
82+
83+
Happy Modeling! 💡

‎llama3-8b-instruct/README.md‎

Lines changed: 87 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,87 @@
1+
# Llama3 8B Instruct
2+
3+
## 📖 Introduction
4+
5+
Meta developed and released the Meta [Llama 3](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks.
6+
7+
| Task Type | Description |
8+
| -------------------------------------------------------- | ------------------------------------------------------------------------------------------- |
9+
| [Chat](https://www.instill.tech/docs/model/ai-task#chat) | A task to generate conversational style text output base on single or multi-modality input. |
10+
11+
## 🔄 Compatibility Matrix
12+
13+
To ensure smooth integration, please refer to the compatibility matrix below. It outlines the compatible versions of the model, [`instill-core`](https://github.com/instill-ai/instill-core), and the [`python-sdk`](https://github.com/instill-ai/python-sdk).
14+
15+
| Model Version | Instill-Core Version | Python-SDK Version |
16+
| ------------- | -------------------- | ------------------ |
17+
| v0.1.0 | >v0.39.0-beta | >0.11.0 |
18+
19+
> **Note:** Always ensure that you are using compatible versions to avoid unexpected issues.
20+
21+
## 🚀 Preparation
22+
23+
Follow [this](../README.md) guide to get your custom model up and running! But before you do that, please read through the following sections to have all the necessary files ready.
24+
25+
#### Install Python SDK
26+
27+
Install the compatible [`python-sdk`](https://github.com/instill-ai/python-sdk) version according to the compatibility matrix:
28+
29+
```bash
30+
pip install instill-sdk=={version}
31+
```
32+
33+
#### Get model weights
34+
35+
To download the fine-tuned model weights, please execute the following command:
36+
37+
```bash
38+
git clone https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct
39+
```
40+
41+
## Test model image
42+
43+
After you've built the model image, and before pushing the model onto any **Instill Core** instance, you can test if the model can be successfully run locally first, by running the following command:
44+
45+
```bash
46+
instill run instill-ai/llama3-8b-instruct -g -i '{"prompt": "hows life?"}'
47+
```
48+
49+
The input payload should strictly follow the the below format
50+
51+
```json
52+
{
53+
"prompt": "..."
54+
}
55+
```
56+
57+
A successful response will return a similar output to that shown below.
58+
59+
```bash
60+
2024-09-11 01:48:37,795.795 INFO [Instill] Starting model image...
61+
2024-09-11 01:48:48,785.785 INFO [Instill] Deploying model...
62+
2024-09-11 01:49:28,613.613 INFO [Instill] Running inference...
63+
2024-09-11 01:49:33,350.350 INFO [Instill] Outputs:
64+
[{'data': {'choices': [{'created': 1725990573,
65+
'finish-reason': 'length',
66+
'index': 0,
67+
'message': {'content': "I'm just an AI, I don't have a "
68+
'life in the classical sense. I '
69+
'exist solely to assist and '
70+
'communicate with users like '
71+
"you. I don't have emotions, "
72+
'experiences, or personal '
73+
"relationships. I'm just a "
74+
'collection of code and data',
75+
'role': 'assistant'}}]}}]
76+
2024-09-11 01:49:37,114.114 INFO [Instill] Done
77+
```
78+
79+
Here is the list of flags supported by `instill run` command
80+
81+
- -t, --tag: tag for the model image, default to `latest`
82+
- -g, --gpu: to pass through GPU from host into container or not, depends on if `gpu` is enabled in the config.
83+
- -i, --input: input in json format
84+
85+
---
86+
87+
Happy Modeling! 💡

0 commit comments

Comments
 (0)