From 9ead7175313ab55f6323dc6b06c378d4335fdbe5 Mon Sep 17 00:00:00 2001
From: aswny <87371411+aswny@users.noreply.github.com>
Date: Thu, 25 Apr 2024 17:19:55 +0000
Subject: [PATCH 1/2] fix Llama models message to prompt conversion in for AWS
 Bedrock provider

---
 litellm/llms/bedrock.py                  |  4 ++++
 litellm/llms/prompt_templates/factory.py | 10 ++++++++++
 2 files changed, 14 insertions(+)

diff --git a/litellm/llms/bedrock.py b/litellm/llms/bedrock.py
index ef6dbfb1b8..149b684724 100644
--- a/litellm/llms/bedrock.py
+++ b/litellm/llms/bedrock.py
@@ -653,6 +653,10 @@ def convert_messages_to_prompt(model, messages, provider, custom_prompt_dict):
         prompt = prompt_factory(
             model=model, messages=messages, custom_llm_provider="bedrock"
         )
+    elif provider == "meta":
+        prompt = prompt_factory(
+            model=model, messages=messages, custom_llm_provider="bedrock"
+        )
     else:
         prompt = ""
         for message in messages:
diff --git a/litellm/llms/prompt_templates/factory.py b/litellm/llms/prompt_templates/factory.py
index a6d1d64386..00c6229c9d 100644
--- a/litellm/llms/prompt_templates/factory.py
+++ b/litellm/llms/prompt_templates/factory.py
@@ -1346,6 +1346,16 @@ def prompt_factory(
                 return anthropic_pt(messages=messages)
         elif "mistral." in model:
             return mistral_instruct_pt(messages=messages)
+        elif "llama2" in model:
+            return llama_2_chat_pt(messages=messages)
+        elif "llama3" in model:
+            return hf_chat_template(
+                model=model,
+                messages=messages,
+                chat_template=known_tokenizer_config[  # type: ignore
+                    "meta-llama/Meta-Llama-3-8B-Instruct"
+                ]["tokenizer"]["chat_template"],
+            )
     elif custom_llm_provider == "perplexity":
         for message in messages:
             message.pop("name", None)

From 781af56f485c0d6b3b33aa71cb34df125e3e7103 Mon Sep 17 00:00:00 2001
From: aswny <87371411+aswny@users.noreply.github.com>
Date: Thu, 25 Apr 2024 17:52:38 +0000
Subject: [PATCH 2/2] check model type chat/instruct to apply template

---
 litellm/llms/prompt_templates/factory.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/litellm/llms/prompt_templates/factory.py b/litellm/llms/prompt_templates/factory.py
index 00c6229c9d..ed719f600d 100644
--- a/litellm/llms/prompt_templates/factory.py
+++ b/litellm/llms/prompt_templates/factory.py
@@ -1346,9 +1346,9 @@ def prompt_factory(
                 return anthropic_pt(messages=messages)
         elif "mistral." in model:
             return mistral_instruct_pt(messages=messages)
-        elif "llama2" in model:
+        elif "llama2" in model and "chat" in model:
             return llama_2_chat_pt(messages=messages)
-        elif "llama3" in model:
+        elif "llama3" in model and "instruct" in model:
             return hf_chat_template(
                 model=model,
                 messages=messages,