File size: 3,162 Bytes
a082853
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2989049
a082853
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
---
license: apache-2.0
tags:
- TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
- abideen/Heimer-dpo-TinyLlama-1.1B
- abideen/Heimer-kto-TinyLlama-1.1B
- Intel/orca_dpo_pairs
language:
- en
datasets:
- Intel/orca_dpo_pairs
library_name: transformers
---

# Heimer-ipo-TinyLlama-1.1B


![image/jpeg](https://cdn-uploads.huggingface.co/production/uploads/64e380b2e12618b261fa6ba0/a7joKICVpqGElN3mh2MpS.jpeg)


# WandB Experiment Tracking

Check out the experiment details in this [report](https://api.wandb.ai/links/zaiinn440/dqlt70dc)


![image/png](https://cdn-uploads.huggingface.co/production/uploads/64e380b2e12618b261fa6ba0/HLhrSUFD1-e6f31F3WGK2.png)

# 🧩 IPO adaptation hyperparameters

## LoRA:

r=8

lora_alpha=16

lora_dropout=0.05

bias="none"

task_type="CAUSAL_LM"

target_modules=['k_proj', 'gate_proj', 'v_proj', 'up_proj', 'q_proj', 'o_proj', 'down_proj']

## Training arguments:

per_device_train_batch_size=2

gradient_accumulation_steps=4

gradient_checkpointing=True

learning_rate=5e-5

lr_scheduler_type="cosine"

max_steps=50

optim="paged_adamw_32bit"

warmup_steps=10

## DPOTrainer:

beta=0.1

max_prompt_length=1024

max_length=1536

loss="ipo"


## 💻 Usage

Here's a [Colab notebook](https://colab.research.google.com/drive/11KEX1LG3nRBoeGR0Iyy-459XllGlLOA9?usp=sharing) to run Heimer-TinyLLama-1.1B in 4-bit precision on a free T4 GPU.

```python
!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "abideen/Heimer-ipo-TinyLlama-1.1B"
messages = [{"role": "user", "content": "Explain what is Data science."}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])
```

"What is Data Science?
A data scientist is an individual who has a passion for data and knowledge of the technology that can be used to help make sense of data. Data scientists are often involved in the development of new software and software platforms, as well as analyzing and interpreting data.
What are the Important components of Data Science?
1. Data: The data is the most important component of a data science project. Data science is the application of data science to make sense of data. Data scientists usually work with data, but data scientists are not necessarily data scientists.
2. Analysis: This is the process of taking data and turning it into something useful.
3. Modeling: The use of machine learning and statistical techniques.
4. Prediction: The prediction of a future event, such as the future market share of a product or the future population of an area.
5. Visualization: Displaying the data in a graphical or interactive format.
6. Statistics: The use of statistical analysis techniques.
What are the Advantages of Data Science?
Data science is the application of data science to make sense of data."