theblackcat102
commited on
Commit
•
1015917
1
Parent(s):
11909cf
Update README.md
Browse files
README.md
CHANGED
@@ -1,8 +1,17 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
4 |
|
5 |
-
#
|
6 |
|
7 |
<!-- Provide a quick summary of what the model is/does. -->
|
8 |
|
@@ -15,49 +24,32 @@ This modelcard aims to be a base template for new models. It has been generated
|
|
15 |
<!-- Provide a longer summary of what this model is. -->
|
16 |
|
17 |
|
18 |
-
|
19 |
-
- **
|
20 |
-
- **
|
21 |
-
- **
|
22 |
-
- **Language(s) (NLP):** [More Information Needed]
|
23 |
-
- **License:** [More Information Needed]
|
24 |
-
- **Finetuned from model [optional]:** [More Information Needed]
|
25 |
|
26 |
## Model Sources [optional]
|
27 |
|
28 |
<!-- Provide the basic links for the model. -->
|
29 |
|
30 |
-
- **Repository:** [
|
31 |
-
- **Paper [optional]:** [More Information Needed]
|
32 |
-
- **Demo [optional]:** [More Information Needed]
|
33 |
|
34 |
# Uses
|
35 |
|
36 |
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
37 |
|
|
|
38 |
## Direct Use
|
39 |
|
40 |
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
41 |
-
|
42 |
-
[More Information Needed]
|
43 |
-
|
44 |
-
## Downstream Use [optional]
|
45 |
-
|
46 |
-
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
47 |
-
|
48 |
-
[More Information Needed]
|
49 |
-
|
50 |
-
## Out-of-Scope Use
|
51 |
-
|
52 |
-
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
53 |
-
|
54 |
-
[More Information Needed]
|
55 |
|
56 |
# Bias, Risks, and Limitations
|
57 |
|
58 |
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
59 |
|
60 |
-
[
|
61 |
|
62 |
## Recommendations
|
63 |
|
@@ -69,7 +61,29 @@ Users (both direct and downstream) should be made aware of the risks, biases and
|
|
69 |
|
70 |
Use the code below to get started with the model.
|
71 |
|
72 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
73 |
|
74 |
# Training Details
|
75 |
|
@@ -77,15 +91,11 @@ Use the code below to get started with the model.
|
|
77 |
|
78 |
<!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
79 |
|
80 |
-
[More Information Needed]
|
81 |
-
|
82 |
## Training Procedure
|
83 |
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
[More Information Needed]
|
89 |
|
90 |
|
91 |
### Training Hyperparameters
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
tags:
|
6 |
+
- sft
|
7 |
+
pipeline_tag: text-generation
|
8 |
+
widget:
|
9 |
+
- text: <prefix>You are a helpful assistant model trained by LAION called Aki</prefix><human>Hi, how are you?<bot>
|
10 |
+
- text: <human>What's the Earth total population<bot>
|
11 |
+
- text: <human>Write a story about future of AI development<bot>
|
12 |
---
|
13 |
|
14 |
+
# Pythia 3B SFT model
|
15 |
|
16 |
<!-- Provide a quick summary of what the model is/does. -->
|
17 |
|
|
|
24 |
<!-- Provide a longer summary of what this model is. -->
|
25 |
|
26 |
|
27 |
+
- **Developed by:** Open Assistant
|
28 |
+
- **Model type:** Pythia
|
29 |
+
- **Language(s) (NLP):** English
|
30 |
+
- **License:** Apache-2.0
|
|
|
|
|
|
|
31 |
|
32 |
## Model Sources [optional]
|
33 |
|
34 |
<!-- Provide the basic links for the model. -->
|
35 |
|
36 |
+
- **Repository:** [Open Assistant](https://github.com/LAION-AI/Open-Assistant)
|
|
|
|
|
37 |
|
38 |
# Uses
|
39 |
|
40 |
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
41 |
|
42 |
+
|
43 |
## Direct Use
|
44 |
|
45 |
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
46 |
+
See the example on the right
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
47 |
|
48 |
# Bias, Risks, and Limitations
|
49 |
|
50 |
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
51 |
|
52 |
+
[just read pythia](https://huggingface.co/EleutherAI/pythia-12b#out-of-scope-use)
|
53 |
|
54 |
## Recommendations
|
55 |
|
|
|
61 |
|
62 |
Use the code below to get started with the model.
|
63 |
|
64 |
+
```python
|
65 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
66 |
+
|
67 |
+
model_name = "theblackcat102/pythia-3b-deduped-sft"
|
68 |
+
|
69 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
70 |
+
model = AutoModelForCausalLM.from_pretrained(model_name).half().eval().cuda()
|
71 |
+
|
72 |
+
input_text = "<human>What's the earth population?<bot>"
|
73 |
+
inputs = tokenizer(input_text, return_tensors="pt", padding=True).to(0)
|
74 |
+
outputs = model.generate(
|
75 |
+
**inputs,
|
76 |
+
early_stopping=True,
|
77 |
+
max_new_tokens=args.max_new_tokens,
|
78 |
+
do_sample=True,
|
79 |
+
top_k=args.top_k,
|
80 |
+
temperature=args.temperature,
|
81 |
+
pad_token_id=tokenizer.eos_token_id,
|
82 |
+
# dialogue_collator.py line 36
|
83 |
+
)
|
84 |
+
output = tokenizer.decode(outputs[0], truncate_before_pattern=[r"\n\n^#", "^'''", "\n\n\n"])
|
85 |
+
print(output)
|
86 |
+
```
|
87 |
|
88 |
# Training Details
|
89 |
|
|
|
91 |
|
92 |
<!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
93 |
|
|
|
|
|
94 |
## Training Procedure
|
95 |
|
96 |
+
```
|
97 |
+
deepspeed trainer_sft.py --configs defaults pythia-3b --deepspeed
|
98 |
+
```
|
|
|
|
|
99 |
|
100 |
|
101 |
### Training Hyperparameters
|