🚀 Synatra-7B-v0.3-RP🐧
Synatra-7B-v0.3-RP is a text - generation model based on the Mistral architecture, offering high - quality text generation capabilities.
🚀 Quick Start
Since the chat_template already contains the instruction format, you can use the following code to start using the model:
from transformers import AutoModelForCausalLM, AutoTokenizer
device = "cuda"
model = AutoModelForCausalLM.from_pretrained("maywell/Synatra-7B-v0.3-RP")
tokenizer = AutoTokenizer.from_pretrained("maywell/Synatra-7B-v0.3-RP")
messages = [
{"role": "user", "content": "바나나는 원래 하얀색이야?"},
]
encodeds = tokenizer.apply_chat_template(messages, return_tensors="pt")
model_inputs = encodeds.to(device)
model.to(device)
generated_ids = model.generate(model_inputs, max_new_tokens=1000, do_sample=True)
decoded = tokenizer.batch_decode(generated_ids)
print(decoded[0])
✨ Features
- Based on the mistralai/Mistral - 7B - Instruct - v0.1 base model.
- Follows the ChatML instruction format.
- The following tasks have been completed:
- Produce an RP - based tuned model.
- Clean the dataset.
- Supplement common sense.
- Planned tasks include improving language understanding and changing the tokenizer.
📦 Installation
The README does not provide specific installation steps, so this section is skipped.
💻 Usage Examples
Basic Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
device = "cuda"
model = AutoModelForCausalLM.from_pretrained("maywell/Synatra-7B-v0.3-RP")
tokenizer = AutoTokenizer.from_pretrained("maywell/Synatra-7B-v0.3-RP")
messages = [
{"role": "user", "content": "바나나는 원래 하얀색이야?"},
]
encodeds = tokenizer.apply_chat_template(messages, return_tensors="pt")
model_inputs = encodeds.to(device)
model.to(device)
generated_ids = model.generate(model_inputs, max_new_tokens=1000, do_sample=True)
decoded = tokenizer.batch_decode(generated_ids)
print(decoded[0])
📚 Documentation
Model Details
Model Benchmark
Ko - LLM - Leaderboard
On Benchmarking...
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
Metric |
Value |
Avg. |
57.38 |
ARC (25 - shot) |
62.2 |
HellaSwag (10 - shot) |
82.29 |
MMLU (5 - shot) |
60.8 |
TruthfulQA (0 - shot) |
52.64 |
Winogrande (5 - shot) |
76.48 |
GSM8K (5 - shot) |
21.15 |
DROP (3 - shot) |
46.06 |
Why It's benchmark score is lower than preview version?
Apparently, the preview model uses the Alpaca Style prompt which has no pre - fix, but ChatML does.
🔧 Technical Details
The README does not provide specific technical details, so this section is skipped.
📄 License
This model is strictly [non - commercial](https://creativecommons.org/licenses/by - nc/4.0/) (cc - by - nc - 4.0) use only.
The "Model" is completely free (i.e., base model, derivates, merges/mixes) to use for non - commercial purposes as long as the included cc - by - nc - 4.0 license in any parent repository, and the non - commercial use statute remains, regardless of other models' licences.
The licence can be changed after a new model is released. If you are to use this model for commercial purposes, contact the author.
⚠️ Important Note
Synatra is a personal project developed with the resources of a single person. If you like the model, you can support the research by buying the author a coffee. If you want to be a sponsor, contact the author on Telegram AlzarTakkarsen.