129 lines
4.0 KiB
Markdown
129 lines
4.0 KiB
Markdown
---
|
|
license: apache-2.0
|
|
---
|
|
|
|
# Synthia-7B-v3.0
|
|
SynthIA-7B-v3.0 (Synthetic Intelligent Agent) is a Mistral-7B model trained with guidance on Orca-2 paper. It has been fine-tuned for instruction following as well as having long-form conversations. SynthIA-3.0 dataset contains the Generarized Tree-of-Thought prompt plus 10 more new long-form system contexts. However, in the training phase the system context was removed as suggested in Orca-2 paper.
|
|
|
|
<br>
|
|
|
|
|
|
|
|
<br>
|
|
|
|
|
|
To evoke generalized Tree of Thought + Chain of Thought reasoning, you may use the following system message:
|
|
```
|
|
Elaborate on the topic using a Tree of Thoughts and backtrack when necessary to construct a clear, cohesive Chain of Thought reasoning. Always answer without hesitation.
|
|
```
|
|
|
|
|
|
## Evaluation
|
|
|
|
We evaluated Synthia-7B-v3.0 on a wide range of tasks using [Language Model Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness) from EleutherAI.
|
|
|
|
Here are the results on metrics used by [HuggingFaceH4 Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard). Section to follow.
|
|
|
|
||||
|
|
|:------:|:--------:|:-------:|
|
|
|**Task**|**Metric**|**Value**|
|
|
|*arc_challenge*|acc_norm||
|
|
|*hellaswag*|acc_norm||
|
|
|*mmlu*|acc_norm||
|
|
|*truthfulqa_mc*|mc2||
|
|
|**Total Average**|-|||
|
|
|
|
<br>
|
|
|
|
## Example Usage
|
|
|
|
### Here is prompt format:
|
|
|
|
```
|
|
SYSTEM: Elaborate on the topic using a Tree of Thoughts and backtrack when necessary to construct a clear, cohesive Chain of Thought reasoning. Always answer without hesitation.
|
|
USER: What is the difference between an Orca, Dolphin and a Seal?
|
|
ASSISTANT:
|
|
```
|
|
|
|
### Below shows a code example on how to use this model:
|
|
|
|
```python
|
|
import torch, json
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
|
|
|
model_path = "migtissera/Synthia-7B-v3.0"
|
|
output_file_path = "./Synthia-7B-conversations.jsonl"
|
|
|
|
model = AutoModelForCausalLM.from_pretrained(
|
|
model_path,
|
|
torch_dtype=torch.float16,
|
|
device_map="auto",
|
|
load_in_8bit=False,
|
|
trust_remote_code=True,
|
|
)
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
|
|
|
|
|
|
def generate_text(instruction):
|
|
tokens = tokenizer.encode(instruction)
|
|
tokens = torch.LongTensor(tokens).unsqueeze(0)
|
|
tokens = tokens.to("cuda")
|
|
|
|
instance = {
|
|
"input_ids": tokens,
|
|
"top_p": 1.0,
|
|
"temperature": 0.75,
|
|
"generate_len": 1024,
|
|
"top_k": 50,
|
|
}
|
|
|
|
length = len(tokens[0])
|
|
with torch.no_grad():
|
|
rest = model.generate(
|
|
input_ids=tokens,
|
|
max_length=length + instance["generate_len"],
|
|
use_cache=True,
|
|
do_sample=True,
|
|
top_p=instance["top_p"],
|
|
temperature=instance["temperature"],
|
|
top_k=instance["top_k"],
|
|
num_return_sequences=1,
|
|
)
|
|
output = rest[0][length:]
|
|
string = tokenizer.decode(output, skip_special_tokens=True)
|
|
answer = string.split("USER:")[0].strip()
|
|
return f"{answer}"
|
|
|
|
|
|
conversation = f"SYSTEM: Elaborate on the topic using a Tree of Thoughts and backtrack when necessary to construct a clear, cohesive Chain of Thought reasoning. Always answer without hesitation."
|
|
|
|
|
|
while True:
|
|
user_input = input("You: ")
|
|
llm_prompt = f"{conversation} \nUSER: {user_input} \nASSISTANT: "
|
|
answer = generate_text(llm_prompt)
|
|
print(answer)
|
|
conversation = f"{llm_prompt}{answer}"
|
|
json_data = {"prompt": user_input, "answer": answer}
|
|
|
|
## Save your conversation
|
|
with open(output_file_path, "a") as output_file:
|
|
output_file.write(json.dumps(json_data) + "\n")
|
|
|
|
```
|
|
|
|
<br>
|
|
|
|
#### Limitations & Biases:
|
|
|
|
While this model aims for accuracy, it can occasionally produce inaccurate or misleading results.
|
|
|
|
Despite diligent efforts in refining the pretraining data, there remains a possibility for the generation of inappropriate, biased, or offensive content.
|
|
|
|
Exercise caution and cross-check information when necessary. This is an uncensored model.
|
|
|
|
|
|
<br>
|
|
|