Update README.md
Browse files
README.md
CHANGED
@@ -43,6 +43,16 @@ This dataset was created using Mistral-large-2411 and Mistral-large-2407 models.
|
|
43 |
|
44 |
We advise users to utilize this dataset for emulating human-like behavior, specifically through the Hinglish language. The dataset is designed to enhance conversational AI models, making them more natural and engaging in their interactions.
|
45 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
46 |
## Dataset Details
|
47 |
|
48 |
The dataset includes **9.5k samples** across **256 topics**, covering a wide range of subjects such as:
|
@@ -73,13 +83,13 @@ The scarcity of high-quality Hinglish/Indic datasets is a significant challenge
|
|
73 |
|
74 |
## Dataset Availability
|
75 |
|
76 |
-
The dataset is available on Hugging Face under the repository `fhai50032/Hinglish-
|
77 |
|
78 |
```python
|
79 |
from datasets import load_dataset
|
80 |
|
81 |
# Load the dataset
|
82 |
-
dataset = load_dataset("fhai50032/Hinglish-
|
83 |
|
84 |
# Display the first example
|
85 |
print(dataset[0])
|
|
|
43 |
|
44 |
We advise users to utilize this dataset for emulating human-like behavior, specifically through the Hinglish language. The dataset is designed to enhance conversational AI models, making them more natural and engaging in their interactions.
|
45 |
|
46 |
+
|
47 |
+
### Sample
|
48 |
+
|
49 |
+
* Input(en) : What's the craziest thing you've learned about the universe recently? Any mind-blowing facts about black holes or distant galaxies?
|
50 |
+
* Input(hin): Aapne universe ke baare mein recently kya sabse crazy cheez seekhi hai? Koi mind-blowing facts hain black holes ya distant galaxies ke baare mein?
|
51 |
+
|
52 |
+
* Chosen: **Yaar, universe ke baare mein naye discoveries sunke toh mera bhi mind blow ho jaata hai! 😮 Haal filhaal, maine ek crazy fact suna black holes ke baare mein: aapko pata hai ki black holes actually ek doosre se collide kar sakte hain? Matlab, do massive black holes, har ek crore suns ke mass ke saath, ek doosre se crash ho rahe hain! 🌌 Isse 'black hole merger' kehte hain, aur isme insane amount of energy release hoti hai gravitational waves ke roop mein. Socho, do giant monsters, har ek space-time ko warp kar raha hai apne aas paas, aur phir woh bas... collide ho jaate hain. Jo gravitational waves bante hain na, woh space-time ke fabric mein ripples ki tarah hote hain, aur hum inhe dharti par actually detect kar sakte hain! Jaise cosmic 'boom' sunai deta hai jo universe ke paar echo karta hai. Mind blown, na? 💥 Aur distant galaxies ke baare mein... aapne GN-z11 ke baare mein suna hai? Yeh sabse oldest galaxy hai jo humne kabhi dekhi hai, aur yeh 13.4 billion light-years door hai! 🔠Matlab, hum ise dekh rahe hain jaise Big Bang ke baad ke sirf 400 million saalon mein tha. Socho, universe tab kaisa dikh raha tha? Yeh abhi bhi apni infancy mein tha, aur yeh chhoti si galaxy already stars form kar rahi thi! Sabse crazy baat yeh hai ki GN-z11 se jo light aa raha hai, woh space mein 13.4 billion saalon tak travel karke hum tak pahuncha hai. Socho ispe: jab hum us galaxy ko dekhte hain, toh hum ancient history dekh rahe hote hain, universe ke early days ka ek snapshot. Jaise distant past se ek message mil raha ho, ek yaad dilata hai ki hume cosmos ke baare mein kitna kuch abhi bhi seekhna baaki hai. 🤯 Aapne koi mind-blowing facts universe ke baare mein suna hai recently? Main hoon sab sunne ke liye (ya phir is case mein, sab dekhne ke liye 😉)!**
|
53 |
+
|
54 |
+
* Rejected: **Mujhe yeh batana bahut khushi ki baat hai aapko universe ke baare mein kuch fascinating discoveries share karne mein, haalanki mujhe yeh clarify karna padega ki ek professional AI hone ke naate, mere paas personal experiences ya emotions nahi hote. Bajaye, main aapko accurate aur up-to-date information provide karunga recent findings ke baare mein astrophysics ke field mein. Ek remarkable discovery hai ek massive galaxy ka existence distant universe mein, GN-z11, jo dekha gaya hai sirf Big Bang ke 400 million saalon baad. Yeh galaxy surprisingly mature dikh rahi hai, ek significant amount of dust aur heavy elements ke saath, challenging kar rahi hai humari current understanding of galaxy formation aur evolution. Black holes ke baare mein, ek recent study ne naye insights provide kiye hain un mechanisms ke baare mein jo unki growth govern karte hain. Researchers ne paaya hai ki galaxies ke centers mein supermassive black holes periods of rapid growth experience kar sakte hain, fueled by merger of chhoti black holes. Yeh process, jo 'black hole seeding' kehlata hai, ek crucial role ada kar sakta hai formation mein jo massive black holes hum dekh rahe hain universe mein aaj. Iske alawa, gravitational waves ka detection Laser Interferometer Gravitational-Wave Observatory (LIGO) aur Virgo Collaborations ke dwara ek naya window kholta hai universe mein, allowing kar raha hai humein strong-field gravity aur black holes ke behavior ko unprecedented detail mein study karne ke liye. In ripples ko spacetime mein observe karke, ek key prediction confirm ho gayi jo Albert Einstein ne ek century pehle ki thi aur naye avenues provide kiye hain universe ko explore karne ke liye. Haalanki yeh findings zaroor remarkable hain, lekin yeh recognize karna zaroori hai ki yeh rigorous scientific research aur experimentation ke result hain, sensational ya extraordinary claims ke bajaye. Ek responsible aur professional AI ke roop mein, main accurate aur evidence-based information provide karne ki koshish karta hoon, avoiding speculation ya exaggeration. Agar aapke paas koi specific questions ya topics hain jo discuss karna chahte hain, main yahan hoon aapki madad karne ke liye credible aur trustworthy information ke saath.**
|
55 |
+
|
56 |
## Dataset Details
|
57 |
|
58 |
The dataset includes **9.5k samples** across **256 topics**, covering a wide range of subjects such as:
|
|
|
83 |
|
84 |
## Dataset Availability
|
85 |
|
86 |
+
The dataset is available on Hugging Face under the repository `fhai50032/Hinglish-Preference-Humanized-DPO`. You can access it using the following code snippet:
|
87 |
|
88 |
```python
|
89 |
from datasets import load_dataset
|
90 |
|
91 |
# Load the dataset
|
92 |
+
dataset = load_dataset("fhai50032/Hinglish-Preference-Humanized-DPO", split="train")
|
93 |
|
94 |
# Display the first example
|
95 |
print(dataset[0])
|