GeoCLIP (code, colab, models)

.gitattributes

@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+GeoCLIP.[[:space:]]Clip-Inspired[[:space:]]Alignment[[:space:]]between[[:space:]]Locations[[:space:]]and[[:space:]]Images[[:space:]]for[[:space:]]Effective[[:space:]]Worldwide[[:space:]]Geo-localization.pdf filter=lfs diff=lfs merge=lfs -text
+GeoClip.[[:space:]]Geometry-Aware[[:space:]]Clipping[[:space:]]for[[:space:]]Differentially[[:space:]]Private[[:space:]]SGD.pdf filter=lfs diff=lfs merge=lfs -text
+Supplementary[[:space:]]for[[:space:]]'GeoCLIP.[[:space:]]Clip-Inspired[[:space:]]Alignment[[:space:]]between[[:space:]]Locations[[:space:]]and[[:space:]]Images[[:space:]]for[[:space:]]Effective[[:space:]]Worldwide[[:space:]]Geo-localization'.pdf filter=lfs diff=lfs merge=lfs -text

GeoCLIP. Clip-Inspired Alignment between Locations and Images for Effective Worldwide Geo-localization.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:527fd1adeaa863c270a0c9872844ddecedb0ddeb722db73d8a9fe1bc5f8256a8
+size 3248321

GeoClip. Geometry-Aware Clipping for Differentially Private SGD.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d0d754ea9b5cb2c8977f66b7cda5f2e218fa47956c52a116d5551aa8759b46f
+size 524315

Supplementary for 'GeoCLIP. Clip-Inspired Alignment between Locations and Images for Effective Worldwide Geo-localization'.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20feb304ad99150b8a4a73dd262f7b171191da1ba93a77452e51ae9a411e6ebc
+size 7881434

code/geo-clip [cem-sirin] +6 -17.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e389616b38d51da3bda1c74b1096e477e2cc4995b1fa3c34fe245618fb0c4d74
+size 85449819

code/geo-clip [selenasun1618] +16 -1.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db8569bdea467a25983c9f5db7481e949feae3a5f6d4164d44b60646ef102cd9
+size 5909705230

code/geo-clip.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3558bfb9b9af0d66b39c77f9e334194541a526556ef6375d4a53580dfc8c5ce
+size 86482350

models/geo-clip/fine_tuned_image_encoder_mlp_weights_08-29-02_48.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:affbfd638cb4eca4f02c4f907b966be5bc8b415f20f13ef0eb27c4d0ced4d97c
+size 3939800

models/geo-clip/fine_tuned_location_encoder_weights_08-29-02_48.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d7bcb338186bbb9f0b17e42264f35cc605bca4ed3ff378e423ad268004d2772
+size 37809053

models/geo-clip/fine_tuned_logit_scale_weights_08-29-02_48.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b48105180ed9d6d75557abb0529e816df9229a82517749e47783140a052707ab
+size 1483

models/geo-clip/image_encoder_mlp_weights.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5342c391eec961796d04d02f2bd670bfa97b34ce32c656d0e0676581103c2c14
+size 3938907

models/geo-clip/location_encoder_weights.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94b80ae3af89fca78539e129dd2929d321247f5ceeeb0a31bc2c31041a253394
+size 37806041

models/geo-clip/logit_scale_weights.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9568e2e09dd19083082bf285b2c0c126867e40dbd275a8ee8c621888b861904e
+size 759

models/geo-clip/source.txt

@@ -0,0 +1 @@
+https://github.com/selenasun1618/geo-clip

models/geoCLIP-tiny-resisc45/.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

models/geoCLIP-tiny-resisc45/README.md

@@ -0,0 +1,29 @@
+
+---
+library_name: transformers
+tags:
+- clip
+- satellite
+- geolocation
+- resisc45
+---
+
+# 🌍 GeoCLIP – Tiny CLIP Trained on RESISC45 (Satellite Scenes)
+
+Fine-tuned Tiny CLIP (≈ 19 M params) from scratch on **RESISC45** satellite scenes.  
+Pairs: *(image, "A satellite image of {label}")*  
+
+## Training Summary
+| Parameter | Value |
+|:--|--:|
+| Steps | 300 (tiny subset) |
+| Optimizer | AdamW (3e-4 lr) |
+| Loss ↓ | 1.49 (final) |
+| Retrieval Top-5 ↑ | ≈ 15 % |
+
+## Visuals
+![Loss curve](loss_curve_20251008_025207.png)  
+![Retrieval bar](retrieval_bar_20251008_025208.png)
+
+---
+Generated 2025-10-08 02:58 UTC

models/geoCLIP-tiny-resisc45/meta.json

@@ -0,0 +1,7 @@
+{
+  "draft": "edt-transformers-project-draft2",
+  "created_at": "2025-10-07T23:56:58.826549+00:00",
+  "subset_size": 512,
+  "steps": 300,
+  "batch": 32
+}

models/geoCLIP-tiny-resisc45/source.txt

@@ -0,0 +1 @@
+https://huggingface.co/paph3285/geoCLIP-tiny-resisc45

models/geoCLIP-tiny-resisc45/tiny_geoclip_tinyset.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c72ebc3780ccde8039a1262dbd828a20bd0e60d44e69de1ed52334dd91e463f
+size 77504316

models/geoclip-large-patch14/.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

models/geoclip-large-patch14/README.md

@@ -0,0 +1,120 @@
+---
+license: mit
+library_name: transformers.js
+---
+
+https://github.com/VicenteVivan/geo-clip with ONNX weights to be compatible with Transformers.js.
+
+## Usage (Transformers.js)
+
+If you haven't already, you can install the [Transformers.js](https://huggingface.co/docs/transformers.js) JavaScript library from [NPM](https://www.npmjs.com/package/@xenova/transformers) using:
+```bash
+npm i @xenova/transformers
+```
+
+**Example:** Perform worldwide image geolocalization.
+```js
+import { AutoModel, AutoProcessor, RawImage, Tensor, dot, softmax } from '@xenova/transformers';
+
+// Load vision and location models
+const model_id = 'Xenova/geoclip-large-patch14';
+const vision_model = await AutoModel.from_pretrained(model_id, {
+    model_file_name: 'vision_model',
+});
+const location_model = await AutoModel.from_pretrained(model_id, {
+    model_file_name: 'location_model',
+    quantized: false,
+});
+
+// Load image processor
+const processor = await AutoProcessor.from_pretrained('openai/clip-vit-large-patch14');
+
+// Read and preprocess image
+const url = 'https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/moraine-lake.png';
+const image = await RawImage.fromURL(url);
+const vision_inputs = await processor(image);
+
+// Compute image embeddings
+const { image_embeds } = await vision_model(vision_inputs);
+const norm_image_embeds = image_embeds.normalize().data;
+
+// Define a list of candidate GPS coordinates
+// https://github.com/VicenteVivan/geo-clip/blob/main/geoclip/model/gps_gallery/coordinates_100K.csv
+const coordinate_data = 'https://huggingface.co/Xenova/geoclip-large-patch14/resolve/main/gps_gallery/coordinates_100K.json';
+const gps_data = await (await fetch(coordinate_data)).json();
+
+// Compute location embeddings and compare to image embeddings
+const coordinate_batch_size = 512;
+const exp_logit_scale = Math.exp(3.681034803390503); // Used for scaling logits
+const scores = [];
+for (let i = 0; i < gps_data.length; i += coordinate_batch_size) {
+    const chunk = gps_data.slice(i, i + coordinate_batch_size);
+
+    const { location_embeds } = await location_model({
+        location: new Tensor('float32', chunk.flat(), [chunk.length, 2])
+    });
+
+    const norm_location_embeds = location_embeds.normalize().tolist();
+    for (const embed of norm_location_embeds) {
+        const score = exp_logit_scale * dot(norm_image_embeds, embed);
+        scores.push(score);
+    }
+}
+
+// Get top predictions
+const top_k = 50;
+const results = softmax(scores)
+    .map((x, i) => [x, i])
+    .sort((a, b) => b[0] - a[0])
+    .slice(0, top_k)
+    .map(([score, index]) => ({ index, gps: gps_data[index], score }));
+
+console.log('=======================');
+console.log('Top 5 GPS Predictions 📍');
+console.log('=======================');
+for (let i = 0; i < 5; ++i) {
+    console.log(results[i]);
+}
+console.log('=======================');
+```
+
+Outputs:
+```
+=======================
+Top 5 GPS Predictions 📍
+=======================
+{
+  index: 75129,
+  gps: [ 51.327447, -116.183509 ],
+  score: 0.06895832493631944
+}
+{
+  index: 5158,
+  gps: [ 51.326401, -116.18263 ],
+  score: 0.06843383108770337
+}
+{
+  index: 77752,
+  gps: [ 51.328198, -116.180934 ],
+  score: 0.06652924543010541
+}
+{
+  index: 32529,
+  gps: [ 51.327809, -116.180334 ],
+  score: 0.065981075526145
+}
+{
+  index: 461,
+  gps: [ 51.322353, -116.18557 ],
+  score: 0.06476605375767822
+}
+=======================
+```
+
+As seen on the map, this is indeed the location of [Moraine Lake](https://en.wikipedia.org/wiki/Moraine_Lake) in Canada.
+
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/ThtHHkOmXEll2tyGV85GY.png)
+
+---
+
+Note: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using [🤗 Optimum](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`).

models/geoclip-large-patch14/config.json

@@ -0,0 +1,94 @@
+{
+  "model_type": "geoclip",
+  "projection_dim": 768,
+  "location_config": {
+    "sigma": [1, 16, 256],
+    "n": 3,
+    "input_size": 2,
+    "encoded_size": 256
+  },
+  "location_config_dict": {},
+  "torch_dtype": "float32",
+  "transformers_version": null,
+  "vision_config": {
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": null,
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "dropout": 0.0,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "quick_gelu",
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "image_size": 224,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-05,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "clip_vision_model",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 16,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_hidden_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "patch_size": 14,
+    "prefix": null,
+    "problem_type": null,
+    "projection_dim" : 768,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": null,
+    "torchscript": false,
+    "transformers_version": "4.16.0.dev0",
+    "use_bfloat16": false
+  },
+  "vision_config_dict": {
+    "hidden_size": 1024,
+    "intermediate_size": 4096,
+    "num_attention_heads": 16,
+    "num_hidden_layers": 24,
+    "patch_size": 14,
+    "projection_dim": 768
+  },
+  "logit_scale": 3.681034803390503
+}

models/geoclip-large-patch14/gps_gallery/coordinates_100K.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:428bce04a09f854bbd4c09b97b7cabccda9cc9d2debc073145e10ccab228979f
+size 800000

models/geoclip-large-patch14/onnx/location_model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6f75032491fd06e9d37d4d0353e5904c96edd49ee2858358644829a408a0269
+size 37808863

models/geoclip-large-patch14/onnx/location_model_bnb4.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d5d2aace73c62dedba19a13557ebe0940e6b310845ec26e05346ac3ffa83a8d
+size 37804052

models/geoclip-large-patch14/onnx/location_model_fp16.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33a69e22b6b91331327f01621ed93c1e022de98e5df9b2303771f48b15cbb645
+size 18914915

models/geoclip-large-patch14/onnx/location_model_int8.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c605968905a634392587673203a75e87b912e61dcb6b8a63d6c94998d7a295f2
+size 9515134

models/geoclip-large-patch14/onnx/location_model_q4.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d931f22d5db007a65cff24f69c725b7278a8b866cba1e14a70ee05107f7c1662
+size 37818536

models/geoclip-large-patch14/onnx/location_model_quantized.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c605968905a634392587673203a75e87b912e61dcb6b8a63d6c94998d7a295f2
+size 9515134

models/geoclip-large-patch14/onnx/location_model_uint8.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd8c60f5d945f7f1f7de595990137837844f82a2d2491287da898ed82841e419
+size 9515146

models/geoclip-large-patch14/onnx/vision_model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be5e445d71789363360e94657263e077ebbc21111526209707c31ff61c2acfaa
+size 1220407251

models/geoclip-large-patch14/onnx/vision_model_bnb4.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b87bc1ac7987b8f4fe74a7582b050eb15862d41b973117ea7be15daf176f3028
+size 179636432

models/geoclip-large-patch14/onnx/vision_model_fp16.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b6d557e7e927adfe4fd3524987ccbfac88e5a4282e9c0d815b2aef5af4c86db
+size 610739002

models/geoclip-large-patch14/onnx/vision_model_int8.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce2a0aa90833e4b6c67a2e06464acce0d70183ae3b8443ac13e4f9d14853023a
+size 306961304

models/geoclip-large-patch14/onnx/vision_model_q4.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56ab09986e2645d40d1f1699cb26440f515df27babd4c283f0851b18eeffc310
+size 198558769

models/geoclip-large-patch14/onnx/vision_model_quantized.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e1507d402874d8b408f46b7dff5076e834fd9fcb3330d7af0727b34632119f4
+size 306961380

models/geoclip-large-patch14/onnx/vision_model_uint8.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e1507d402874d8b408f46b7dff5076e834fd9fcb3330d7af0727b34632119f4
+size 306961380

models/geoclip-large-patch14/preprocessor_config.json

@@ -0,0 +1,19 @@
+{
+  "crop_size": 224,
+  "do_center_crop": true,
+  "do_normalize": true,
+  "do_resize": true,
+  "feature_extractor_type": "CLIPFeatureExtractor",
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "resample": 3,
+  "size": 224
+}

models/geoclip-large-patch14/quantize_config.json

@@ -0,0 +1,189 @@
+{
+    "fp16": {},
+    "q8": {
+        "per_model_config": {
+            "vision_model": {
+                "op_types": [
+                    "Add",
+                    "Cast",
+                    "Concat",
+                    "Constant",
+                    "ConstantOfShape",
+                    "Conv",
+                    "Div",
+                    "Equal",
+                    "Expand",
+                    "Gather",
+                    "Gemm",
+                    "MatMul",
+                    "Mul",
+                    "Pow",
+                    "ReduceMean",
+                    "Relu",
+                    "Reshape",
+                    "Shape",
+                    "Sigmoid",
+                    "Slice",
+                    "Softmax",
+                    "Sqrt",
+                    "Sub",
+                    "Transpose",
+                    "Unsqueeze",
+                    "Where"
+                ],
+                "weight_type": "QUInt8"
+            },
+            "location_model": {
+                "op_types": [
+                    "Add",
+                    "Asin",
+                    "Cast",
+                    "Concat",
+                    "Constant",
+                    "ConstantOfShape",
+                    "Cos",
+                    "Div",
+                    "Gather",
+                    "Gemm",
+                    "MatMul",
+                    "Mul",
+                    "Pow",
+                    "Relu",
+                    "Shape",
+                    "Sin",
+                    "Unsqueeze"
+                ],
+                "weight_type": "QInt8"
+            }
+        },
+        "per_channel": false,
+        "reduce_range": false
+    },
+    "int8": {
+        "per_model_config": {
+            "vision_model": {
+                "op_types": [
+                    "Add",
+                    "Cast",
+                    "Concat",
+                    "Constant",
+                    "ConstantOfShape",
+                    "Conv",
+                    "Div",
+                    "Equal",
+                    "Expand",
+                    "Gather",
+                    "Gemm",
+                    "MatMul",
+                    "Mul",
+                    "Pow",
+                    "ReduceMean",
+                    "Relu",
+                    "Reshape",
+                    "Shape",
+                    "Sigmoid",
+                    "Slice",
+                    "Softmax",
+                    "Sqrt",
+                    "Sub",
+                    "Transpose",
+                    "Unsqueeze",
+                    "Where"
+                ],
+                "weight_type": "QInt8"
+            },
+            "location_model": {
+                "op_types": [
+                    "Add",
+                    "Asin",
+                    "Cast",
+                    "Concat",
+                    "Constant",
+                    "ConstantOfShape",
+                    "Cos",
+                    "Div",
+                    "Gather",
+                    "Gemm",
+                    "MatMul",
+                    "Mul",
+                    "Pow",
+                    "Relu",
+                    "Shape",
+                    "Sin",
+                    "Unsqueeze"
+                ],
+                "weight_type": "QInt8"
+            }
+        },
+        "per_channel": false,
+        "reduce_range": false
+    },
+    "uint8": {
+        "per_model_config": {
+            "vision_model": {
+                "op_types": [
+                    "Add",
+                    "Cast",
+                    "Concat",
+                    "Constant",
+                    "ConstantOfShape",
+                    "Conv",
+                    "Div",
+                    "Equal",
+                    "Expand",
+                    "Gather",
+                    "Gemm",
+                    "MatMul",
+                    "Mul",
+                    "Pow",
+                    "ReduceMean",
+                    "Relu",
+                    "Reshape",
+                    "Shape",
+                    "Sigmoid",
+                    "Slice",
+                    "Softmax",
+                    "Sqrt",
+                    "Sub",
+                    "Transpose",
+                    "Unsqueeze",
+                    "Where"
+                ],
+                "weight_type": "QUInt8"
+            },
+            "location_model": {
+                "op_types": [
+                    "Add",
+                    "Asin",
+                    "Cast",
+                    "Concat",
+                    "Constant",
+                    "ConstantOfShape",
+                    "Cos",
+                    "Div",
+                    "Gather",
+                    "Gemm",
+                    "MatMul",
+                    "Mul",
+                    "Pow",
+                    "Relu",
+                    "Shape",
+                    "Sin",
+                    "Unsqueeze"
+                ],
+                "weight_type": "QUInt8"
+            }
+        },
+        "per_channel": false,
+        "reduce_range": false
+    },
+    "q4": {
+        "block_size": 32,
+        "is_symmetric": true,
+        "accuracy_level": null
+    },
+    "bnb4": {
+        "block_size": 64,
+        "quant_type": 1
+    }
+}

models/geoclip-large-patch14/source.txt

@@ -0,0 +1 @@
+https://huggingface.co/Xenova/geoclip-large-patch14

models/geoclip-location-encoder/.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

models/geoclip-location-encoder/README.md

@@ -0,0 +1,9 @@
+---
+tags:
+- model_hub_mixin
+- pytorch_model_hub_mixin
+---
+
+This model has been pushed to the Hub using the [PytorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) integration:
+- Library: [More Information Needed]
+- Docs: [More Information Needed]

models/geoclip-location-encoder/config.json

@@ -0,0 +1,11 @@
+{
+  "dim": 512,
+  "encoded_size": 256,
+  "input_size": 2,
+  "sigma": [
+    1,
+    16,
+    256
+  ],
+  "transformers_version": "4.45.2"
+}

models/geoclip-location-encoder/model.py

@@ -0,0 +1,133 @@
+import torch
+import torch.nn as nn
+from typing import Optional
+from torch import Tensor
+import numpy as np
+from huggingface_hub import PyTorchModelHubMixin
+
+# Constants
+A1 = 1.340264
+A2 = -0.081106
+A3 = 0.000893
+A4 = 0.003796
+SF = 66.50336
+
+@torch.jit.script
+def gaussian_encoding(
+        v: Tensor,
+        b: Tensor) -> Tensor:
+    r"""Computes :math:`\gamma(\mathbf{v}) = (\cos{2 \pi \mathbf{B} \mathbf{v}} , \sin{2 \pi \mathbf{B} \mathbf{v}})`
+
+    Args:
+        v (Tensor): input tensor of shape :math:`(N, *, \text{input_size})`
+        b (Tensor): projection matrix of shape :math:`(\text{encoded_layer_size}, \text{input_size})`
+
+    Returns:
+        Tensor: mapped tensor of shape :math:`(N, *, 2 \cdot \text{encoded_layer_size})`
+
+    See :class:`~rff.layers.GaussianEncoding` for more details.
+    """
+    vp = 2 * np.pi * v @ b.T
+    return torch.cat((torch.cos(vp), torch.sin(vp)), dim=-1)
+
+
+def sample_b(sigma: float, size: tuple) -> Tensor:
+    r"""Matrix of size :attr:`size` sampled from from :math:`\mathcal{N}(0, \sigma^2)`
+
+    Args:
+        sigma (float): standard deviation
+        size (tuple): size of the matrix sampled
+
+    See :class:`~rff.layers.GaussianEncoding` for more details
+    """
+    return torch.randn(size) * sigma
+
+class GaussianEncoding(nn.Module):
+    """Layer for mapping coordinates using random Fourier features"""
+
+    def __init__(self, sigma: Optional[float] = None,
+                 input_size: Optional[float] = None,
+                 encoded_size: Optional[float] = None,
+                 b: Optional[Tensor] = None):
+        r"""
+        Args:
+            sigma (Optional[float]): standard deviation
+            input_size (Optional[float]): the number of input dimensions
+            encoded_size (Optional[float]): the number of dimensions the `b` matrix maps to
+            b (Optional[Tensor], optional): Optionally specify a :attr:`b` matrix already sampled
+        Raises:
+            ValueError:
+                If :attr:`b` is provided and one of :attr:`sigma`, :attr:`input_size`,
+                or :attr:`encoded_size` is provided. If :attr:`b` is not provided and one of
+                :attr:`sigma`, :attr:`input_size`, or :attr:`encoded_size` is not provided.
+        """
+        super().__init__()
+        if b is None:
+            if sigma is None or input_size is None or encoded_size is None:
+                raise ValueError(
+                    'Arguments "sigma," "input_size," and "encoded_size" are required.')
+
+            b = sample_b(sigma, (encoded_size, input_size))
+        elif sigma is not None or input_size is not None or encoded_size is not None:
+            raise ValueError('Only specify the "b" argument when using it.')
+        self.b = nn.parameter.Parameter(b, requires_grad=False)
+
+    def forward(self, v: Tensor) -> Tensor:
+        r"""Computes :math:`\gamma(\mathbf{v}) = (\cos{2 \pi \mathbf{B} \mathbf{v}} , \sin{2 \pi \mathbf{B} \mathbf{v}})`
+
+        Args:
+            v (Tensor): input tensor of shape :math:`(N, *, \text{input_size})`
+
+        Returns:
+            Tensor: Tensor mapping using random fourier features of shape :math:`(N, *, 2 \cdot \text{encoded_size})`
+        """
+        return gaussian_encoding(v, self.b)
+
+def equal_earth_projection(L):
+    latitude = L[:, 0]
+    longitude = L[:, 1]
+    latitude_rad = torch.deg2rad(latitude)
+    longitude_rad = torch.deg2rad(longitude)
+    sin_theta = (torch.sqrt(torch.tensor(3.0)) / 2) * torch.sin(latitude_rad)
+    theta = torch.asin(sin_theta)
+    denominator = 3 * (9 * A4 * theta**8 + 7 * A3 * theta**6 + 3 * A2 * theta**2 + A1)
+    x = (2 * torch.sqrt(torch.tensor(3.0)) * longitude_rad * torch.cos(theta)) / denominator
+    y = A4 * theta**9 + A3 * theta**7 + A2 * theta**3 + A1 * theta
+    return (torch.stack((x, y), dim=1) * SF) / 180
+
+class LocationEncoderCapsule(nn.Module):
+    def __init__(self, sigma):
+        super(LocationEncoderCapsule, self).__init__()
+        rff_encoding = GaussianEncoding(sigma=sigma, input_size=2, encoded_size=256)
+        self.km = sigma
+        self.capsule = nn.Sequential(rff_encoding,
+                                     nn.Linear(512, 1024),
+                                     nn.ReLU(),
+                                     nn.Linear(1024, 1024),
+                                     nn.ReLU(),
+                                     nn.Linear(1024, 1024),
+                                     nn.ReLU())
+        self.head = nn.Sequential(nn.Linear(1024, 512))
+
+    def forward(self, x):
+        x = self.capsule(x)
+        x = self.head(x)
+        return x
+
+class LocationEncoder(nn.Module, PyTorchModelHubMixin):
+    def __init__(self):
+        super(LocationEncoder, self).__init__()
+        self.sigma = [2**0, 2**4, 2**8]
+        self.n = len(self.sigma)
+
+        for i, s in enumerate(self.sigma):
+            self.add_module('LocEnc' + str(i), LocationEncoderCapsule(sigma=s))
+
+    def forward(self, location):
+        location = equal_earth_projection(location)
+        location_features = torch.zeros(location.shape[0], 512).to(location.device)
+
+        for i in range(self.n):
+            location_features += self._modules['LocEnc' + str(i)](location)
+        
+        return location_features

models/geoclip-location-encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02b8fc91b9857c6db2bd7f8808df2f34a25d2e279d4b5ae7447510d31bf75913
+size 37800376