Ben Graville commited on
Commit
2bee70c
·
verified ·
1 Parent(s): a88aa55

embeddinggemma-300m Core AI fp32 static (seq 256), rebuilt on Linux box 2026-08-17; producer coreai-core 1.0.0b2; tokenizer bit-identical to mlboydaisuke@d9a60a18, main.mlirb +262 B (conversion is not byte-deterministic)

Browse files
.gitattributes CHANGED
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ model/embeddinggemma-300m_float32_static.aimodel/main.mlirb filter=lfs diff=lfs merge=lfs -text
37
+ model/tokenizer/tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: gemma
3
+ base_model: google/embeddinggemma-300m
4
+ library_name: coreai
5
+ tags:
6
+ - coreai
7
+ - aimodel
8
+ - sentence-embedding
9
+ - on-device
10
+ - apple-silicon
11
+ pipeline_tag: sentence-similarity
12
+ extra_gated_heading: Access EmbeddingGemma on Hugging Face
13
+ extra_gated_description: >-
14
+ This artifact is a derivative of google/embeddinggemma-300m and is governed by
15
+ the Gemma Terms of Use, the Gemma Prohibited Use Policy and the Gemma license.
16
+ ---
17
+
18
+ # embeddinggemma-300m — Core AI (.aimodel)
19
+
20
+ `google/embeddinggemma-300m` exported as a single static fp32 Core AI graph
21
+ (transformer -> mean pooling -> dense stack -> L2 normalize, all in-graph).
22
+
23
+ | | |
24
+ |---|---|
25
+ | Recipe | `coreai-model-zoo/conversion/export_embeddinggemma.py --dtype float32 --seq-len 256` |
26
+ | Recipe index | `models/qwen3-embedding/recipe.toml` -> `["embeddinggemma-300m"]` (`status = "verified"`) |
27
+ | Upstream bundle | `mlboydaisuke/embeddinggemma-300m-CoreAI` |
28
+ | Producer | `coreai-core 1.0.0b2` (coreai-torch 0.4.1, torch 2.9.0) |
29
+ | I/O | `input_ids [1,256] int32`, `attention_mask [1,256] int32` -> `embedding [1,768]` fp32, L2-normalized |
30
+
31
+ Built on Linux x86_64 (conversion only). **Runtime parity is not measured here** —
32
+ Core AI cannot execute on Linux; the on-device parity gate runs on macOS against
33
+ `model/reference.json`.
34
+
35
+ ## License
36
+
37
+ Derivative of EmbeddingGemma; use is subject to the
38
+ [Gemma Terms of Use](https://ai.google.dev/gemma/terms) and the
39
+ [Gemma Prohibited Use Policy](https://ai.google.dev/gemma/prohibited_use_policy).
model/embeddinggemma-300m_float32_static.aimodel/main.hash ADDED
@@ -0,0 +1 @@
 
 
1
+ �������٫;��rj�HH�#��E��D�ݤ�
model/embeddinggemma-300m_float32_static.aimodel/main.mlirb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f02b286f1f2e09fd9ab3bcffb726ac74848ec8e23baa545f28cd6448ddda4a9
3
+ size 1244157766
model/embeddinggemma-300m_float32_static.aimodel/metadata.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "author" : "Google DeepMind",
3
+ "assetVersion" : "2.0",
4
+ "creationDate" : "20260817T201458Z",
5
+ "producer" : "coreai-core 1.0.0b2",
6
+ "description" : "EmbeddingGemma-300m text embedding model (mean pooling + dense projection, L2-normalized 768-d). Source: https:\/\/huggingface.co\/google\/embeddinggemma-300m",
7
+ "license" : "Gemma"
8
+ }
model/reference.json ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "google/embeddinggemma-300m",
3
+ "seq_len": 256,
4
+ "dtype": "float32",
5
+ "prompts": {
6
+ "query": "task: search result | query: ",
7
+ "document": "title: none | text: ",
8
+ "BitextMining": "task: search result | query: ",
9
+ "Clustering": "task: clustering | query: ",
10
+ "Classification": "task: classification | query: ",
11
+ "InstructionRetrieval": "task: code retrieval | query: ",
12
+ "MultilabelClassification": "task: classification | query: ",
13
+ "PairClassification": "task: sentence similarity | query: ",
14
+ "Reranking": "task: search result | query: ",
15
+ "Retrieval": "task: search result | query: ",
16
+ "Retrieval-query": "task: search result | query: ",
17
+ "Retrieval-document": "title: none | text: ",
18
+ "STS": "task: sentence similarity | query: ",
19
+ "Summarization": "task: summarization | query: "
20
+ },
21
+ "texts": {
22
+ "query_bike": {
23
+ "kind": "query",
24
+ "text": "red bicycle parked at the beach"
25
+ },
26
+ "query_capital": {
27
+ "kind": "query",
28
+ "text": "what is the capital of Japan"
29
+ },
30
+ "doc_bike": {
31
+ "kind": "document",
32
+ "text": "A crimson bike leaning against a palm tree by the sea."
33
+ },
34
+ "doc_tokyo": {
35
+ "kind": "document",
36
+ "text": "Tokyo is the capital and largest city of Japan."
37
+ }
38
+ },
39
+ "cosines": {
40
+ "query_bike|query_capital": 0.1459198147058487,
41
+ "doc_bike|query_bike": 0.6273255348205566,
42
+ "doc_bike|query_capital": 0.03804422914981842,
43
+ "doc_bike|doc_tokyo": 0.20749269425868988,
44
+ "doc_tokyo|query_bike": 0.08935077488422394,
45
+ "doc_tokyo|query_capital": 0.5779669880867004
46
+ }
47
+ }
model/tokenizer/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bcb89366c4787d27671d021cdc006defcd1cb53b02b04213ae8dd47c009c7f71
3
+ size 33385272
model/tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "boi_token": "<start_of_image>",
4
+ "bos_token": "<bos>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eoi_token": "<end_of_image>",
7
+ "eos_token": "<eos>",
8
+ "image_token": "<image_soft_token>",
9
+ "is_local": false,
10
+ "local_files_only": false,
11
+ "mask_token": "<mask>",
12
+ "model_max_length": 2048,
13
+ "model_specific_special_tokens": {
14
+ "boi_token": "<start_of_image>",
15
+ "eoi_token": "<end_of_image>",
16
+ "image_token": "<image_soft_token>"
17
+ },
18
+ "pad_token": "<pad>",
19
+ "padding_side": "right",
20
+ "sp_model_kwargs": null,
21
+ "spaces_between_special_tokens": false,
22
+ "tokenizer_class": "GemmaTokenizer",
23
+ "unk_token": "<unk>",
24
+ "use_default_system_prompt": false
25
+ }