Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feedback fixes round 1 #333

Merged
merged 12 commits into from
Jan 14, 2025
Merged

Feedback fixes round 1 #333

merged 12 commits into from
Jan 14, 2025

Conversation

trentfowlercohere
Copy link
Contributor

@trentfowlercohere trentfowlercohere commented Jan 7, 2025

This PR introduces several changes to the documentation of Cohere's models and their applications. The updates aim to provide a more comprehensive and accurate overview of the models' capabilities and usage.

  • Updated Model Descriptions: The PR modifies the titles and descriptions of several model-related pages, including the "Models Overview" page, the "Embed Model" page, and the "Rerank Model" page. The new titles and descriptions offer a clearer and more detailed introduction to Cohere's models, emphasizing their specific applications and features.
  • Enhanced Model Details: The PR adds more information about the rerank-v3.5 and rerank-v3.0 models, specifying that these models consider both the query and the document when calculating against the context length limit. It also clarifies that the query can account for up to half of the full context length, and if it exceeds 2048 tokens, it will be truncated to that length, leaving the remaining space for the document(s).
  • Improved Model Access Information: The PR updates the content related to accessing Cohere's models, emphasizing the Cohere SDK as the primary method. It also provides a link to create a trial or production API key and includes information about rate limits.
  • Clarified Model Training: The PR clarifies that users can train a custom model, allowing them to create models tailored to their specific use cases and trained on their data.
  • LLM Limitations and Prompt Truncation: The PR includes a section about the limitations of LLMs, specifically their text input capacity. It introduces the prompt_truncation parameter of the Chat API, which automatically breaks up documents into smaller chunks, reranks them based on relevance, and drops the least relevant documents to stay within the model's context length limit.

Copy link

github-actions bot commented Jan 7, 2025

Copy link

github-actions bot commented Jan 7, 2025

Copy link

github-actions bot commented Jan 7, 2025

Copy link

github-actions bot commented Jan 8, 2025

Copy link

github-actions bot commented Jan 8, 2025

@billytrend-cohere billytrend-cohere self-requested a review January 9, 2025 16:55
trentfowlercohere and others added 2 commits January 13, 2025 16:57
Co-authored-by: billytrend-cohere <[email protected]>
Signed-off-by: trentfowlercohere <[email protected]>
Copy link

@trentfowlercohere trentfowlercohere merged commit 047f2cb into main Jan 14, 2025
3 checks passed
@trentfowlercohere trentfowlercohere deleted the feedback-fixes-round-1 branch January 14, 2025 16:36
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants