Instructions to use google/gemma-4-E4B-it with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use google/gemma-4-E4B-it with Transformers:
# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("google/gemma-4-E4B-it") model = AutoModelForMultimodalLM.from_pretrained("google/gemma-4-E4B-it", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- AMD Developer Cloud
fix: scope preserve_thinking to tool-call turns only
Browse files- chat_template.jinja +1 -1
chat_template.jinja
CHANGED
|
@@ -230,7 +230,7 @@
|
|
| 230 |
|
| 231 |
{#- Render reasoning/reasoning_content as thinking channel -#}
|
| 232 |
{%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}
|
| 233 |
-
{%- set thinking_gate = (loop.index0 > ns_turn.last_user_idx) or preserve_thinking -%}
|
| 234 |
{%- if thinking_text and thinking_gate -%}
|
| 235 |
{{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
|
| 236 |
{%- endif -%}
|
|
|
|
| 230 |
|
| 231 |
{#- Render reasoning/reasoning_content as thinking channel -#}
|
| 232 |
{%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}
|
| 233 |
+
{%- set thinking_gate = (loop.index0 > ns_turn.last_user_idx) or (preserve_thinking and message.get('tool_calls')) -%}
|
| 234 |
{%- if thinking_text and thinking_gate -%}
|
| 235 |
{{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
|
| 236 |
{%- endif -%}
|