From 87c6868cff8a94d9f008c9d75466c8f1d43202f4 Mon Sep 17 00:00:00 2001 From: Thibault Genaitay Date: Fri, 15 Nov 2024 14:38:06 +0100 Subject: [PATCH 1/4] feat(ifr): extended fc support --- .../reference-content/function-calling-support.mdx | 1 + 1 file changed, 1 insertion(+) diff --git a/ai-data/managed-inference/reference-content/function-calling-support.mdx b/ai-data/managed-inference/reference-content/function-calling-support.mdx index 5d153ccf45..97cd953605 100644 --- a/ai-data/managed-inference/reference-content/function-calling-support.mdx +++ b/ai-data/managed-inference/reference-content/function-calling-support.mdx @@ -29,6 +29,7 @@ The following models in Scaleway's Managed Inference library can call tools as p * meta/llama-3.1-70b-instruct * mistral/mistral-7b-instruct-v0.3 * mistral/mistral-nemo-instruct-2407 +* nvidia/llama-3.1-nemotron-70b-instruct ## Understanding function calling From 6c393539197145be4af56894eb9e78303dcfb42c Mon Sep 17 00:00:00 2001 From: Thibault Genaitay Date: Fri, 15 Nov 2024 15:23:09 +0100 Subject: [PATCH 2/4] feat(ifr): adding nemotron to ref content --- .../function-calling-support.mdx | 2 +- .../llama-3.1-nemotron-70b-instruct.mdx | 81 +++++++++++++++++++ menu/navigation.json | 4 + 3 files changed, 86 insertions(+), 1 deletion(-) create mode 100644 ai-data/managed-inference/reference-content/llama-3.1-nemotron-70b-instruct.mdx diff --git a/ai-data/managed-inference/reference-content/function-calling-support.mdx b/ai-data/managed-inference/reference-content/function-calling-support.mdx index 97cd953605..0b085a32af 100644 --- a/ai-data/managed-inference/reference-content/function-calling-support.mdx +++ b/ai-data/managed-inference/reference-content/function-calling-support.mdx @@ -7,7 +7,7 @@ content: paragraph: Function calling allows models to connect to external tools. tags: dates: - validation: 2024-10-25 + validation: 2024-11-15 posted: 2024-10-25 categories: - ai-data diff --git a/ai-data/managed-inference/reference-content/llama-3.1-nemotron-70b-instruct.mdx b/ai-data/managed-inference/reference-content/llama-3.1-nemotron-70b-instruct.mdx new file mode 100644 index 0000000000..71175f290a --- /dev/null +++ b/ai-data/managed-inference/reference-content/llama-3.1-nemotron-70b-instruct.mdx @@ -0,0 +1,81 @@ +--- +meta: + title: Understanding the Llama-3.1-Nemotron-70b-instruct model + description: Deploy your own secure Llama-3.1-Nemotron-70b-instruct model with Scaleway Managed Inference. Privacy-focused, fully managed. +content: + h1: Understanding the Llama-3.1-Nemotron-70b-instruct model + paragraph: This page provides information on the Llama-3.1-Nemotron-70b-instruct model +tags: +dates: + validation: 2024-11-15 + posted: 2024-11-15 +categories: + - ai-data +--- + +## Model overview + +| Attribute | Details | +|-----------------|------------------------------------| +| Provider | [Nvidia](https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Instruct) | +| License | [Llama 3.1 community](https://llama.meta.com/llama3_1/license/) | | +| Compatible Instances | H100 (FP8), H100-2 (FP8) | +| Context Length | up to 128k tokens | + +## Model names + +```bash +meta/llama-3.1-nemotron-70b-instruct:fp8 +``` + +## Compatible Instances + +| Instance type | Max context length | +| ------------- |-------------| +| H100 | 128k (FP8) | +| H100-2 | 128k (FP8) | + +## Model introduction + +Introduced October 14, 2024, NVIDIA's Nemotron 70B Instruct is a specialized version of the Llama 3.1 model designed to follow complex instructions. +NVIDIA employed Reinforcement Learning from Human Feedback (RLHF) to fine-tune the model’s ability to generate relevant and informative responses. + +## Why is it useful? + +- As of October 2024, Llama 3.1 Nemotron 70B has achieved top rankings in multiple automatic alignment benchmarks. It boasts an overall score of 94.1 on RewardBench, with specific scores of 97.5 for chat performance and 98.1 in reasoning tasks. +- Just like with the original Llama 3.1, this model brings a context window up to 128K tokens and [supports tool calling](/ai-data/managed-inference/reference-content/function-calling-support). +- With 70 billion parameters, this model is highly capable of generating sophisticated, human-like responses in a wide range of applications, from casual chatbots to complex technical systems. + +## How to use it + +### Sending Managed Inference requests + +To perform inference tasks with your Llama-3.1-Nemotron-70b-instruct deployed at Scaleway, use the following command: + +```bash +curl -s \ +-H "Authorization: Bearer " \ +-H "Content-Type: application/json" \ +--request POST \ +--url "https://.ifr.fr-par.scaleway.com/v1/chat/completions" \ +--data '{"model":"meta/llama-3.1-nemotron-70b-instruct:fp8", "messages":[{"role": "user","content": "There is a llama in my garden, what should I do?"}], "max_tokens": 500, "temperature": 0.7, "stream": false}' +``` + +Make sure to replace `` and `` with your actual [IAM API key](/identity-and-access-management/iam/how-to/create-api-keys/) and the Deployment UUID you are targeting. + + + The model name allows Scaleway to put your prompts in the expected format. + + + + Ensure that the `messages` array is properly formatted with roles (system, user, assistant) and content. + + +### Receiving Inference responses + +Upon sending the HTTP request to the public or private endpoints exposed by the server, you will receive inference responses from the managed Managed Inference server. +Process the output data according to your application's needs. The response will contain the output generated by the LLM model based on the input provided in the request. + + + Despite efforts for accuracy, the possibility of generated text containing inaccuracies or [hallucinations](/ai-data/managed-inference/concepts/#hallucinations) exists. Always verify the content generated independently. + \ No newline at end of file diff --git a/menu/navigation.json b/menu/navigation.json index 82904cc59b..6aa8046a3b 100644 --- a/menu/navigation.json +++ b/menu/navigation.json @@ -611,6 +611,10 @@ "label": "Llama-3.1-70b-instruct model", "slug": "llama-3.1-70b-instruct" }, + { + "label": "Llama-3.1-nemotron-70b-instruct model", + "slug": "llama-3.1-nemotron-70b-instruct" + }, { "label": "Mistral-nemo-instruct-2407 model", "slug": "mistral-nemo-instruct-2407" From be83b24d82c36961d1c32982e7d738abb2eb4561 Mon Sep 17 00:00:00 2001 From: Thibault Genaitay Date: Fri, 15 Nov 2024 15:24:08 +0100 Subject: [PATCH 3/4] fix(ifr): mixtral switched from FP16 to BF16 --- .../reference-content/mixtral-8x7b-instruct-v0.1.mdx | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/ai-data/managed-inference/reference-content/mixtral-8x7b-instruct-v0.1.mdx b/ai-data/managed-inference/reference-content/mixtral-8x7b-instruct-v0.1.mdx index db496f4b33..7744171ca4 100644 --- a/ai-data/managed-inference/reference-content/mixtral-8x7b-instruct-v0.1.mdx +++ b/ai-data/managed-inference/reference-content/mixtral-8x7b-instruct-v0.1.mdx @@ -18,14 +18,14 @@ categories: | Attribute | Details | |-----------------|------------------------------------| | Provider | [Mistral](https://mistral.ai/technology/#models) | -| Compatible Instances | H100 (FP8) - H100-2 (FP16) | +| Compatible Instances | H100 (FP8) - H100-2 (BF16) | | Context size | 32k tokens | ## Model names ```bash mistral/mixtral-8x7b-instruct-v0.1:fp8 -mistral/mixtral-8x7b-instruct-v0.1:fp16 +mistral/mixtral-8x7b-instruct-v0.1:bf16 ``` ## Compatible Instances @@ -33,7 +33,7 @@ mistral/mixtral-8x7b-instruct-v0.1:fp16 | Instance type | Max context length | | ------------- |-------------| | H100 | 32k (FP8) -| H100-2 | 32k (FP16) +| H100-2 | 32k (BF16) ## Model introduction From 89f9dc58300a06c24c3b6a39af3fe38e9b408acb Mon Sep 17 00:00:00 2001 From: Thibault Genaitay Date: Mon, 18 Nov 2024 10:39:24 +0100 Subject: [PATCH 4/4] feat(ifr): updated support --- ai-data/generative-apis/how-to/use-function-calling.mdx | 3 ++- .../reference-content/function-calling-support.mdx | 3 ++- .../reference-content/llama-3.1-nemotron-70b-instruct.mdx | 2 +- 3 files changed, 5 insertions(+), 3 deletions(-) diff --git a/ai-data/generative-apis/how-to/use-function-calling.mdx b/ai-data/generative-apis/how-to/use-function-calling.mdx index 4f58e4ac66..0d7dc4f08c 100644 --- a/ai-data/generative-apis/how-to/use-function-calling.mdx +++ b/ai-data/generative-apis/how-to/use-function-calling.mdx @@ -7,7 +7,7 @@ content: paragraph: Learn how to enhance AI applications by integrating external tools using Scaleway's Chat Completions API service. tags: chat-completions-api dates: - validation: 2024-10-30 + validation: 2024-11-18 posted: 2024-09-24 --- @@ -28,6 +28,7 @@ Function calling allows a large language model (LLM) to interact with external t * llama-3.1-8b-instruct * llama-3.1-70b-instruct * mistral-nemo-instruct-2407 +* pixtral-12b-2409 ## Understanding function calling diff --git a/ai-data/managed-inference/reference-content/function-calling-support.mdx b/ai-data/managed-inference/reference-content/function-calling-support.mdx index 0b085a32af..757a297e97 100644 --- a/ai-data/managed-inference/reference-content/function-calling-support.mdx +++ b/ai-data/managed-inference/reference-content/function-calling-support.mdx @@ -7,7 +7,7 @@ content: paragraph: Function calling allows models to connect to external tools. tags: dates: - validation: 2024-11-15 + validation: 2024-11-18 posted: 2024-10-25 categories: - ai-data @@ -29,6 +29,7 @@ The following models in Scaleway's Managed Inference library can call tools as p * meta/llama-3.1-70b-instruct * mistral/mistral-7b-instruct-v0.3 * mistral/mistral-nemo-instruct-2407 +* mistral/pixtral-12b-2409 * nvidia/llama-3.1-nemotron-70b-instruct ## Understanding function calling diff --git a/ai-data/managed-inference/reference-content/llama-3.1-nemotron-70b-instruct.mdx b/ai-data/managed-inference/reference-content/llama-3.1-nemotron-70b-instruct.mdx index 71175f290a..52b4e1eed6 100644 --- a/ai-data/managed-inference/reference-content/llama-3.1-nemotron-70b-instruct.mdx +++ b/ai-data/managed-inference/reference-content/llama-3.1-nemotron-70b-instruct.mdx @@ -32,7 +32,7 @@ meta/llama-3.1-nemotron-70b-instruct:fp8 | Instance type | Max context length | | ------------- |-------------| -| H100 | 128k (FP8) | +| H100 | 16k (FP8) | | H100-2 | 128k (FP8) | ## Model introduction