feat: add warnings in documentation about provider transparency

genai-impact · Sep 23, 2024 · d33a8c5 · d33a8c5
1 parent 0771090
commit d33a8c5
Show file tree

Hide file tree

Showing 7 changed files with 29 additions and 0 deletions.
diff --git a/docs/tutorial/providers/anthropic.md b/docs/tutorial/providers/anthropic.md
@@ -1,5 +1,9 @@
 # Anthropic
 
+!!! danger "Lack of transparency"
+
+    Anthropic does not disclose any information about model architecture and inference infrastrucure. Thus, the environmental impacts are estimated with a very low precision.
+
 This guide focuses on the integration of :seedling: **EcoLogits** with the [Anthropic official python client :octicons-link-external-16:](https://github.com/anthropics/anthropic-sdk-python).
 
 Official links:

diff --git a/docs/tutorial/providers/cohere.md b/docs/tutorial/providers/cohere.md
@@ -1,5 +1,9 @@
 # Cohere
 
+!!! warning "Lack of transparency"
+
+    Majority of models released by Cohere are open-weights, but there is no information on the inference infrastructure. Thus, the environmental impacts are estimated with a lower precision.
+
 This guide focuses on the integration of :seedling: **EcoLogits** with the [Cohere official python client :octicons-link-external-16:](https://github.com/cohere-ai/cohere-python).
 
 Official links:

diff --git a/docs/tutorial/providers/google.md b/docs/tutorial/providers/google.md
@@ -1,5 +1,9 @@
 # Google Gemini
 
+!!! danger "Lack of transparency"
+
+    Google does not disclose any information about model architecture and inference infrastrucure. Thus, the environmental impacts are estimated with a very low precision.
+
 This guide focuses on the integration of :seedling: **EcoLogits** with the [Google Gemini official python client :octicons-link-external-16:](https://github.com/google-gemini/generative-ai-python).
 
 Official links:

diff --git a/docs/tutorial/providers/huggingface_hub.md b/docs/tutorial/providers/huggingface_hub.md
@@ -1,5 +1,9 @@
 # Hugging Face Hub
 
+!!! info "Lack of transparency"
+
+    All models available on Hugging Face inference endpoints are open-weights, but there is not information on the inference infrastructure. Thus, the environmental impacts are estimated with a lower precision.
+
 This guide focuses on the integration of :seedling: **EcoLogits** with the [Hugging Face Hub official python client :octicons-link-external-16:](https://github.com/huggingface/huggingface_hub).
 
 Official links:

diff --git a/docs/tutorial/providers/litellm.md b/docs/tutorial/providers/litellm.md
@@ -1,5 +1,9 @@
 # LiteLLM
 
+!!! warning "Lack of transparency"
+
+    Depending on the provider, the level of transparency varies. Thus, the environmental impacts can be estimated with a low precision.
+
 This guide focuses on the integration of :seedling: **EcoLogits** with the [LiteLLM official Python client :octicons-link-external-16:](https://github.com/BerriAI/litellm).
 
 Official links:

diff --git a/docs/tutorial/providers/mistralai.md b/docs/tutorial/providers/mistralai.md
@@ -4,6 +4,11 @@
 
     Mistral AI python client with version `<1.0.0` will no longer be supported by EcoLogits. See official [migration guide from v0 to v1 :octicons-link-external-16:](https://github.com/mistralai/client-python/blob/main/MIGRATION.md). 
 
+!!! warning "Lack of transparency"
+
+    Some models released by Mistral AI are not open-weights, plus there is no information on the inference infrastructure. Thus, the environmental impacts are estimated with a lower precision.
+
+
 This guide focuses on the integration of :seedling: **EcoLogits** with the [Mistral AI official python client :octicons-link-external-16:](https://github.com/mistralai/client-python).
 
 Official links:

diff --git a/docs/tutorial/providers/openai.md b/docs/tutorial/providers/openai.md
@@ -1,5 +1,9 @@
 # OpenAI
 
+!!! danger "Lack of transparency"
+
+    OpenAI does not disclose any information about model architecture and inference infrastrucure. Thus, the environmental impacts are estimated with a very low precision.
+
 This guide focuses on the integration of :seedling: **EcoLogits** with the [OpenAI official python client :octicons-link-external-16:](https://github.com/openai/openai-python).
 
 Official links: