As part of this, we've updated LangSmith to support rendering of Tools and Tool Calls in traces for various models, including @AnthropicAI
!
CODE
-
LangSmith Updates Tool Rendering Support for Anthropic Models
By
–
-
HuggingFace Trainer DistributedDataParallel Gradient Bug Fix
By
–
ok so, it was a gradient bug. let me explain with the huggingface trainer, when training on multiple GPUs, sometimes the model is stored as an nn.Module, but sometimes it’s wrapped in this DistributedDataParallel thing I was inadvertently forwarding on the nn.Module in
-
DeepSpeed optimization successfully implemented for AI training
By
–
@winglian wrangled deepspeed till it worked
-
Fine-tuned Mixtral 8x22B Inference Optimization Guide
By
–
Fine-tuned Mixtral 8x22B is ready. Curious what the easiest way to do inference is? Spinning up some GPUs for vLLM right now, is there something better?
-
Prompting Mixtral-8x22B: Effective Techniques Thread
By
–
Man I forgot how much more intensive prompting base models is All of this, just to shorten some text. Would anyone find value in a quick thread on how to prompt Mixtral-8x22B effectively?
-
AI Agent Devin Converts Word Documents to LaTeX PDFs
By
–
AGI achieved. (OK, of course not, but a reminder that I shouldn't assume AI can't do something) "Devin, give me a PDF of a beautiful LaTex version of this Word document. Add at least one illustration. Add a title page with my name, look up my affiliation" Not perfect, but cute
-
HashiCorp Accused of Copying MPL-2.0 Licensed Code
By
–
"The code in question can be clearly shown to have been copied from older code under the MPL-2.0 license. HashiCorp seems to have copied the same code itself when they implemented their version of this feature." Pretty convincing summary right there.
-
Standardized Tool Calling Across Model Providers
By
–
Standardized tool calling **tl;dr;** we've made some big improvements to working with tools across model providers! A new `tool_calls` field directly on messages with the same schema for all providers. A new `ToolCallingAgent`. Supports any model with tool calling
-
LangSmith Evaluations: Leveraging Prompt Hub for LLM Development
By
–
LangSmith Evaluations: Using the prompt hub Evaluations can accelerate LLM app development, but it can be challenging to get started. We've kicked off a new video series focused on evaluations in LangSmith. This is the 8th video in our series that shows how to perform
-
Training Single Models with Cross-Entropy Loss
By
–
thanks for sharing — seems like this is some odd student-teacher setting though; i'm just training a single model with cross-entropy