Update README.md
Browse files
README.md
CHANGED
@@ -1,15 +1,12 @@
|
|
1 |
---
|
2 |
-
license: other
|
3 |
license_name: openmdw
|
4 |
license_link: LICENSE
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
<a href="https://github.com/ByteDance-Seed/Seed-X-7B/blob/main/LICENSE.openmdw">
|
12 |
-
<img src="https://img.shields.io/badge/License-OpenMDW-yellow"></a>
|
13 |
|
14 |
## Introduction
|
15 |
We are excited to introduce **Seed-X**, a powerful series of open-source multilingual translation language models, including an instruction model, a reinforcement learning model, and a reward model. It pushes the boundaries of translation capabilities within 7 billion parameters.
|
@@ -45,9 +42,6 @@ This repo contains the **Seed-X-Instruct** model, with the following features:
|
|
45 |
Here is a simple example demonstrating how to load the model and perform translation using ```vllm```
|
46 |
```python
|
47 |
from vllm import LLM, SamplingParams
|
48 |
-
|
49 |
-
model_path = "./ByteDance-Seed/Seed-X-Instruct-7B"
|
50 |
-
|
51 |
model = LLM(model=model_path,
|
52 |
max_num_seqs=512,
|
53 |
tensor_parallel_size=8,
|
@@ -55,17 +49,10 @@ model = LLM(model=model_path,
|
|
55 |
gpu_memory_utilization=0.95)
|
56 |
|
57 |
messages = [
|
58 |
-
"Translate the following English sentence
|
59 |
-
"Translate the following English sentence
|
60 |
]
|
61 |
|
62 |
-
# Sampling
|
63 |
-
decoding_params = SamplingParams(temperature=0,
|
64 |
-
max_tokens=512,
|
65 |
-
skip_special_tokens=True)
|
66 |
-
# Beam Search
|
67 |
-
decoding_params = BeamSearchParams(beam_width=4,
|
68 |
-
max_tokens=512)
|
69 |
|
70 |
results = model.generate(messages, decoding_params)
|
71 |
responses = [res.outputs[0].text.strip() for res in results]
|
@@ -73,23 +60,4 @@ responses = [res.outputs[0].text.strip() for res in results]
|
|
73 |
print(responses)
|
74 |
```
|
75 |
## Evaluation
|
76 |
-
We evaluated Seed-X on a diverse set
|
77 |
-

|
78 |
-
For detailed benchmark results and analysis, please refer to our [Technical Report](https://github.com/ByteDance-Seed/Seed-X-7B/blob/main/Technical_Report.pdf).
|
79 |
-
|
80 |
-
## License
|
81 |
-
This project is licensed under OpenMDW. See the [LICENSE](https://github.com/ByteDance-Seed/Seed-X-7B/blob/main/LICENSE.openmdw) file for details.
|
82 |
-
|
83 |
-
## Citation
|
84 |
-
<!--If you find Seed-X useful for your research and applications, feel free to give us a star ⭐ or cite us using:
|
85 |
-
```bibtex
|
86 |
-
@Article{XXX,
|
87 |
-
title={XXXXXXXXXXX},
|
88 |
-
author={XXX,XXX,XXX,XXX},
|
89 |
-
year={2025},
|
90 |
-
eprint={XXXX.XXXXX},
|
91 |
-
archivePrefix={arXiv},
|
92 |
-
primaryClass={cs.XX}
|
93 |
-
}
|
94 |
-
```-->
|
95 |
-
We will soon publish our technical report on Arxiv.
|
|
|
1 |
---
|
|
|
2 |
license_name: openmdw
|
3 |
license_link: LICENSE
|
4 |
+
datasets:
|
5 |
+
- fka/awesome
|
6 |
+
metrics:
|
7 |
+
- accuracy
|
8 |
+
- character
|
9 |
+
pipeline_tag: text-classification
|
|
|
|
|
10 |
|
11 |
## Introduction
|
12 |
We are excited to introduce **Seed-X**, a powerful series of open-source multilingual translation language models, including an instruction model, a reinforcement learning model, and a reward model. It pushes the boundaries of translation capabilities within 7 billion parameters.
|
|
|
42 |
Here is a simple example demonstrating how to load the model and perform translation using ```vllm```
|
43 |
```python
|
44 |
from vllm import LLM, SamplingParams
|
|
|
|
|
|
|
45 |
model = LLM(model=model_path,
|
46 |
max_num_seqs=512,
|
47 |
tensor_parallel_size=8,
|
|
|
49 |
gpu_memory_utilization=0.95)
|
50 |
|
51 |
messages = [
|
52 |
+
"Translate the following English sentence :\nMay the force be with you <zh>", # without CoT
|
53 |
+
"Translate the following English sentence and explain it in detail:\nMay the force be with you <zh>" # with CoT
|
54 |
]
|
55 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
56 |
|
57 |
results = model.generate(messages, decoding_params)
|
58 |
responses = [res.outputs[0].text.strip() for res in results]
|
|
|
60 |
print(responses)
|
61 |
```
|
62 |
## Evaluation
|
63 |
+
We evaluated Seed-X on a diverse set
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|