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:
# pip install -U transformers accelerate # 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
Commit History
-
Add response_template to tokenizer_config.json (#39)
- Add response_template to tokenizer_config.json (7797a9a0e4c5893fe88bb8c04f2797e81270fd52) - Update README.md (a7408c3fd3178b448b312cc91a63ad52e933682a) Co-authored-by: Matthew Carrigan <Rocketknight1@users.noreply.huggingface.co>ee0ef60
-
Update README (#42)
- Update README (c57189f3f1902dab8cd2c608e1ab805da64e70b5)fa62d88 -
fix: chat template — null handling, reasoning preservation, turn-tag balance, input validation (#36)
- fix: chat template — null handling, reasoning preservation, turn-tag balance, input validation (050f48ca2b5965e1a67da7e5217a9cd5243d674e) - fix: restore model turn + thinking cue after tool responses (0cc2d28d3e62a7e112a32e275002955f6731c34a) - fix: prevent extra <turn|> when assistant has content + tool_calls + continuation (77a7d3b5fce4207ec7eeeb09a196b24eb63b1ddb) - fix: revert add_generation_prompt regression + preserve_thinking default (bc413dace821bf29efe8ce19c538b230eede114d) - fix: render thinking channel regardless of tool_calls presence (9098ed86cdaee8a24a9bd81c636eb3bf9ab7a0f1) - fix: update stale comment — thinking channel no longer tool-call-only (74cdce6d7964027b46551724db394ab6aa3143d6) - fix: scope preserve_thinking to tool-call turns only (f3667db1aaa288e20f0a9a7e7662516e73e91a4b) - fix: scope preserve_thinking to tool-call turns, remove extra newline in continuation turns (25f079acde96439dbacbfcb4ea3bed0741a9d4ee) - Add canonical header to chat template (ff861a90880…6ea8ce9
-
Add Technical Report
Verified signaturea4c2d58
-
Update README.md
Verified signaturefee6332
-
Update README.md
Verified signaturecd0bc28
-
fix(chat_template): Emit multimodal placeholders in tool response content-parts (#30)
- fix(chat_template): Emit multimodal placeholders in tool response content-parts (a6bf4a78e8412d6b3fb3d9d2f8863b32cd2dfa5d) Co-authored-by: Harshal Janjani <harshaljanjani@users.noreply.huggingface.co>d6436b3
-
Fix lineage (#31)
- Fix lineage (ddfaaa0605dc3786d04a594ac96172d8785b61ed) Co-authored-by: Omar Sanseviero <osanseviero@users.noreply.huggingface.co>3555bdd
-
fix(chat_template): update SI and tool call handling (#28)
- fix(chat_template): update SI and tool call handling (13793714b1df6c8635dfb6f3c6bb96f94b851cd2) Co-authored-by: Douglas Reid <dougreid@users.noreply.huggingface.co>c53e9d3
-
Update response schema (#16)
- Update response schema (1d9dc276ebaec3b1a92712a47a0a1e54e38a5b84) Co-authored-by: Matthew Carrigan <Rocketknight1@users.noreply.huggingface.co>83df0a8 -
fix: prefix-preserving tool calls + dialog compliance (#14)
- fix: prefix-preserving tool calls + dialog compliance (860044107cc8ff99b34b642972fe294338ca0dca) Co-authored-by: Luciano Martins <lucianommartins@users.noreply.huggingface.co>ab4d202
-
Preparing for release!
Co-authored-by: Rocketknight1 <Rocketknight1@users.noreply.huggingface.co> Co-authored-by: RyanMullins <RyanMullins@users.noreply.huggingface.co> Co-authored-by: pcuenq <pcuenq@users.noreply.huggingface.co> Co-authored-by: bebechien <bebechien@users.noreply.huggingface.co> Co-authored-by: osanseviero <osanseviero@users.noreply.huggingface.co> Co-authored-by: douglas reid <douglas reid@users.noreply.huggingface.co>292a7e2