multilingual upgrade upload of gemma-3n

README.md

@@ -3,12 +3,11 @@ license: gemma
 library_name: transformers
 pipeline_tag: image-text-to-text
 extra_gated_heading: Access Gemma on Hugging Face
-extra_gated_prompt: >-
-  To access Gemma on Hugging Face, you’re required to review and agree to
-  Google’s usage license. To do this, please ensure you’re logged in to Hugging
+extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and
+  agree to Google’s usage license. To do this, please ensure you’re logged in to Hugging
   Face and click below. Requests are processed immediately.
 extra_gated_button_content: Acknowledge license
-base_model: google/gemma-3n-E4B
+base_model: google/gemma-3n-E4B-it
 tags:
 - automatic-speech-recognition
 - automatic-speech-translation
@@ -17,16 +16,18 @@ tags:
 ---
 
 > [!Note]
-> This repository corresponds to the launch version of Gemma 3n E4B IT (Instruct), to be used with Hugging Face `transformers`,
+> This repository corresponds to the launch version of Gemma 3n E2B IT (Instruct), to be used with Hugging Face `transformers`,
 > supporting text, audio, and vision (image and video) inputs.
 > 
 > Gemma 3n models have multiple architecture innovations:
->  * They are available in two sizes based on [effective parameters](https://ai.google.dev/gemma/docs/gemma-3n#parameters). While the raw parameter count of this model is 8B, the architecture design allows the model to be run with a memory footprint comparable to a traditional 4B model by offloading low-utilization matrices from the accelerator.
->  * They use a MatFormer architecture that allows nesting sub-models within the E4B model. We provide one sub-model (an [E2B](https://huggingface.co/google/gemma-3n-E2B-it)), or you can access a spectrum of custom-sized models using the [Mix-and-Match method](https://goo.gle/gemma3n-matformer-lab).
+>  * They are available in two sizes based on [effective parameters](https://ai.google.dev/gemma/docs/gemma-3n#parameters). While the raw parameter count of this model is 6B, the architecture design allows the model to be run with a memory footprint comparable to a traditional 2B model by offloading low-utilization matrices from the accelerator.
+>  * They use a MatFormer architecture that allows nesting sub-models within the [E4B model](https://huggingface.co/google/gemma-3n-E4B-it). We provide one sub-model (this model repository), or you can access a spectrum of custom-sized models using the [Mix-and-Match method](https://goo.gle/gemma3n-matformer-lab).
 >
 > Learn more about these techniques in the [technical blog post](https://developers.googleblog.com/en/introducing-gemma-3n-developer-guide)
 > and the [Gemma documentation](https://ai.google.dev/gemma/docs/gemma-3n). 
 
+
+
 # Gemma 3n model card
 
 **Model Page**: [Gemma 3n](https://ai.google.dev/gemma/docs/gemma-3n)
@@ -101,7 +102,7 @@ import torch
 
 pipe = pipeline(
     "image-text-to-text",
-    model="google/gemma-3n-e4b-it",
+    model="google/gemma-3n-e2b-it",
     device="cuda",
     torch_dtype=torch.bfloat16,
 )
@@ -140,9 +141,9 @@ from PIL import Image
 import requests
 import torch
 
-model_id = "google/gemma-3n-e4b-it"
+model_id = "google/gemma-3n-e2b-it"
 
-model = Gemma3nForConditionalGeneration.from_pretrained(model_id, device_map="auto", torch_dtype=torch.bfloat16,).eval()
+model = Gemma3nForConditionalGeneration.from_pretrained(model_id, device="cuda", torch_dtype=torch.bfloat16,).eval()
 
 processor = AutoProcessor.from_pretrained(model_id)
 
@@ -166,7 +167,7 @@ inputs = processor.apply_chat_template(
     tokenize=True,
     return_dict=True,
     return_tensors="pt",
-).to(model.device)
+).to(model.device, dtype=torch.bfloat16)
 
 input_len = inputs["input_ids"].shape[-1]
 
@@ -526,4 +527,4 @@ development compared to similarly sized models.
 
 Using the benchmark evaluation metrics described in this document, these models
 have shown to provide superior performance to other, comparably-sized open model
-alternatives.
+alternatives.

config.json

@@ -90,11 +90,6 @@
       0.0,
       0.0,
       0.0,
-      0.0,
-      0.0,
-      0.0,
-      0.0,
-      0.0,
       0.0
     ],
     "altup_active_idx": 0,
@@ -110,41 +105,36 @@
     "hidden_size_per_layer_input": 256,
     "initializer_range": 0.02,
     "intermediate_size": [
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384,
-      16384
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192,
+      8192
     ],
     "laurel_rank": 64,
     "layer_types": [
@@ -177,19 +167,14 @@
       "sliding_attention",
       "sliding_attention",
       "sliding_attention",
-      "full_attention",
-      "sliding_attention",
-      "sliding_attention",
-      "sliding_attention",
-      "sliding_attention",
       "full_attention"
     ],
     "max_position_embeddings": 32768,
     "model_type": "gemma3n_text",
     "num_attention_heads": 8,
-    "num_hidden_layers": 35,
+    "num_hidden_layers": 30,
     "num_key_value_heads": 2,
-    "num_kv_shared_layers": 15,
+    "num_kv_shared_layers": 10,
     "rms_norm_eps": 1e-06,
     "rope_local_base_freq": 10000.0,
     "rope_scaling": null,

model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1af26e1fd61af0dc067252c907bf52900c7cd5864893e29970e6ea87320322a6
+size 3077103824

model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f53ec36e9b34a1dda547103f7371eaf4dcce40a9e85ef2a04dfa12f30e1146ff
+size 4981242176

model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28e11f2029b8d13aa3fde2a948808baa01d8bee0fd184c4faedbd09065e7fd0b
+size 2820739840