codeShare commited on
Commit
e1a7e79
1 Parent(s): 1a30b74

Upload sd_token_similarity_calculator.ipynb

Browse files
Google Colab Notebooks/sd_token_similarity_calculator.ipynb CHANGED
@@ -2882,10 +2882,10 @@
2882
  ],
2883
  "metadata": {
2884
  "id": "rUXQ73IbonHY",
2885
- "outputId": "0be95ebc-c6b0-4141-d299-d1e33c7a2e00",
2886
  "colab": {
2887
  "base_uri": "https://localhost:8080/"
2888
- }
 
2889
  },
2890
  "execution_count": 1,
2891
  "outputs": [
@@ -3033,10 +3033,10 @@
3033
  ],
3034
  "metadata": {
3035
  "id": "ZMG4CThUAmwW",
3036
- "outputId": "d5ec8719-a0d7-4bff-f87c-379fed3a5d20",
3037
  "colab": {
3038
  "base_uri": "https://localhost:8080/"
3039
- }
 
3040
  },
3041
  "execution_count": 2,
3042
  "outputs": [
@@ -3857,18 +3857,17 @@
3857
  " text_features = text_encodings[f'{index}']\n",
3858
  "\n",
3859
  " torch.matmul(text_features, image_features.t()) * logit_scale\n",
3860
- " sims[index] = torch.dot(text_features, image_features)\n",
3861
  " if name_NEG != '':\n",
3862
  " torch.matmul(text_features_NEG, image_features.t()) * logit_scale\n",
3863
- " sims[index] = sims[index] - neg_strength* torch.dot(text_features_NEG, image_features)\n",
3864
  " if image_NEG != '':\n",
3865
- " sims[index] = sims[index] - neg_strength* torch.dot(image_features, image_features_NEG)\n",
3866
  "#-------#\n",
3867
  "sorted , indices = torch.sort(sims,dim=0 , descending=True)"
3868
  ],
3869
  "metadata": {
3870
  "id": "rebogpoyOG8k",
3871
- "outputId": "c2a9083b-3b49-4693-bb17-4b829182462f",
3872
  "colab": {
3873
  "base_uri": "https://localhost:8080/",
3874
  "height": 273,
@@ -3962,7 +3961,8 @@
3962
  "9b8be4e7d91a47c2852ca202528744cf",
3963
  "78b25863b7cc4400abf53a04e1f67bb2"
3964
  ]
3965
- }
 
3966
  },
3967
  "execution_count": 4,
3968
  "outputs": [
@@ -4127,10 +4127,10 @@
4127
  ],
4128
  "metadata": {
4129
  "id": "JkzncP8SgKtS",
4130
- "outputId": "d47b9b06-982a-457e-ef50-60664ba6c11f",
4131
  "colab": {
4132
  "base_uri": "https://localhost:8080/"
4133
- }
 
4134
  },
4135
  "execution_count": 6,
4136
  "outputs": [
 
2882
  ],
2883
  "metadata": {
2884
  "id": "rUXQ73IbonHY",
 
2885
  "colab": {
2886
  "base_uri": "https://localhost:8080/"
2887
+ },
2888
+ "outputId": "0be95ebc-c6b0-4141-d299-d1e33c7a2e00"
2889
  },
2890
  "execution_count": 1,
2891
  "outputs": [
 
3033
  ],
3034
  "metadata": {
3035
  "id": "ZMG4CThUAmwW",
 
3036
  "colab": {
3037
  "base_uri": "https://localhost:8080/"
3038
+ },
3039
+ "outputId": "d5ec8719-a0d7-4bff-f87c-379fed3a5d20"
3040
  },
3041
  "execution_count": 2,
3042
  "outputs": [
 
3857
  " text_features = text_encodings[f'{index}']\n",
3858
  "\n",
3859
  " torch.matmul(text_features, image_features.t()) * logit_scale\n",
3860
+ " sims[index] = torch.nn.functional.cosine_similarity(text_features, image_features)\n",
3861
  " if name_NEG != '':\n",
3862
  " torch.matmul(text_features_NEG, image_features.t()) * logit_scale\n",
3863
+ " sims[index] = sims[index] - neg_strength* torch.nn.functional.cosine_similarity(text_features_NEG, image_features)\n",
3864
  " if image_NEG != '':\n",
3865
+ " sims[index] = sims[index] - neg_strength* torch.nn.functional.cosine_similarity(image_features, image_features_NEG)\n",
3866
  "#-------#\n",
3867
  "sorted , indices = torch.sort(sims,dim=0 , descending=True)"
3868
  ],
3869
  "metadata": {
3870
  "id": "rebogpoyOG8k",
 
3871
  "colab": {
3872
  "base_uri": "https://localhost:8080/",
3873
  "height": 273,
 
3961
  "9b8be4e7d91a47c2852ca202528744cf",
3962
  "78b25863b7cc4400abf53a04e1f67bb2"
3963
  ]
3964
+ },
3965
+ "outputId": "c2a9083b-3b49-4693-bb17-4b829182462f"
3966
  },
3967
  "execution_count": 4,
3968
  "outputs": [
 
4127
  ],
4128
  "metadata": {
4129
  "id": "JkzncP8SgKtS",
 
4130
  "colab": {
4131
  "base_uri": "https://localhost:8080/"
4132
+ },
4133
+ "outputId": "d47b9b06-982a-457e-ef50-60664ba6c11f"
4134
  },
4135
  "execution_count": 6,
4136
  "outputs": [