https://txt.cohere.com/command-r/
https://huggingface.co/CohereForAI/c4ai-command-r-v01
1. RAG Performance
On multiple datasets, it far exceeds the Mixtral MoE model. By using their own embeddings and reranking, it significantly outperforms open-source models.
2. Tool Capabilities
The tool capabilities are slightly better than Mixtral and significantly outperform GPT-3.5.
3. Multilingual Capabilities
Supports English, French, Spanish, Italian, German, Brazilian Portuguese, Japanese, Korean, Simplified Chinese, and Arabic.
4. Long Text Capability
The needle-in-a-haystack test can achieve fully green results.
5. License
CC-BY-NC, non-commercial use only.
Additional Information
from transformers import AutoTokenizer
model_id = "CohereForAI/c4ai-command-r-v01"
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
# define conversation input:
conversation = [
{"role": "user", "content": "Whats the biggest penguin in the world?"}
]
# Define tools available for the model to use:
tools = [
{
"name": "internet_search",
"description": "Returns a list of relevant document snippets for a textual query retrieved from the internet",
"parameter_definitions": {
"query": {
"description": "Query to search the internet with",
"type": 'str',
"required": True
}
}
},
{
'name': "directly_answer",
"description": "Calls a standard (un-augmented) AI chatbot to generate a response given the conversation history",
'parameter_definitions': {}
}
]
# render the tool use prompt as a string:
tool_use_prompt = tokenizer.apply_tool_use_template(
conversation,
tools=tools,
tokenize=False,
add_generation_prompt=True,
)
print(tool_use_prompt)