Fix README.md typo

FlagEmbedding/llm_dense_retriever/README.md

@@ -11,10 +11,10 @@
 - [x] Checkpoint
 - [x] Training Data
 - [x] Training Code
+- [x] Technical Report
 - [ ] Evaluation Pipeline
-- [ ] Technical Report
 
-We will release the technical report for **BGE-EN-ICL** in the future.
+The technical report for **BGE-EN-ICL** can be found in [Making Text Embedders Few-Shot Learners](https://arxiv.org/abs/2409.15700)
 
 ## Environment
 ```bash
@@ -39,7 +39,7 @@ pip install flash-attn --no-build-isolation
 
 | Data                                                        | Introduction                                                 |
 | ------------------------------------------------------------ | ------------------------------------------------------------ |
-| [e5-data](https://huggingface.co/datasets/cfli/bge-e5data) | Public data identical to [e5-mistral](https://huggingface.co/intfloat/e5-mistral-7b-instruct) |
+| [public-data](https://huggingface.co/datasets/cfli/bge-e5data) | Public data identical to [e5-mistral](https://huggingface.co/intfloat/e5-mistral-7b-instruct) |
 | [full-data](https://huggingface.co/datasets/cfli/bge-full-data) | The full dataset we used for training |
 
 ## Usage 
@@ -219,13 +219,13 @@ If you find this repository useful, please give us a star ⭐.
 To cite our work:
 
 ```
-@misc{li2023makinglargelanguagemodels,
-      title={Making Large Language Models A Better Foundation For Dense Retrieval}, 
-      author={Chaofan Li and Zheng Liu and Shitao Xiao and Yingxia Shao},
-      year={2023},
-      eprint={2312.15503},
+@misc{li2024makingtextembeddersfewshot,
+      title={Making Text Embedders Few-Shot Learners}, 
+      author={Chaofan Li and MingHao Qin and Shitao Xiao and Jianlyu Chen and Kun Luo and Yingxia Shao and Defu Lian and Zheng Liu},
+      year={2024},
+      eprint={2409.15700},
       archivePrefix={arXiv},
-      primaryClass={cs.CL},
-      url={https://arxiv.org/abs/2312.15503}, 
+      primaryClass={cs.IR},
+      url={https://arxiv.org/abs/2409.15700}, 
 }
 ```

FlagEmbedding/visual/modeling.py

@@ -100,6 +100,7 @@ def __init__(self,
             self.to(self.device)
         else:
             self.device = torch.device('cpu')
+        self.dtype = next(bge.parameters()).dtype
 
     def load_model(self, model_weight):
         self.load_state_dict(torch.load(model_weight, map_location='cpu'))
@@ -191,7 +192,7 @@ def encode_text(self, texts):
         token_type_ids = torch.zeros(input_shape, dtype=torch.long, device=device)
 
         head_mask = [None] * self.depth
-        extended_attention_mask: torch.Tensor = self.get_extended_attention_mask(attention_mask, input_shape)
+        extended_attention_mask: torch.Tensor = self.get_extended_attention_mask(attention_mask, input_shape).to(self.dtype)
 
         embedding_output = self.bge_embeddings(
             input_ids=input_ids,
@@ -270,7 +271,7 @@ def encode_mm(self, images:torch.Tensor, texts):
         prom_img_input_shape = prompt_img_embedding.size()
 
         head_mask = [None] * self.depth
-        extended_attention_mask: torch.Tensor = self.get_extended_attention_mask(prom_img_attention_mask, prom_img_input_shape).to(prompt_img_embedding.dtype)
+        extended_attention_mask: torch.Tensor = self.get_extended_attention_mask(prom_img_attention_mask, prom_img_input_shape).to(self.dtype)
 
 
         encoder_outputs = self.bge_encoder(

README.md

@@ -43,7 +43,7 @@ FlagEmbedding focuses on retrieval-augmented LLMs, consisting of the following p
 - **Benchmark**: [C-MTEB](https://github.com/FlagOpen/FlagEmbedding/tree/master/C_MTEB), [AIR-Bench](https://github.com/AIR-Bench/AIR-Bench), [MLVU](https://github.com/JUNJIE99/MLVU)
 
 ## News
-- 9/10/2014: Introducing **MemoRAG**, a step forward towards RAG 2.0 on top of memory-inspired knowledge discovery (repo: https://github.com/qhjqhj00/MemoRAG, paper: https://arxiv.org/pdf/2409.05591v1) :fire:
+- 9/10/2024: Introducing **MemoRAG**, a step forward towards RAG 2.0 on top of memory-inspired knowledge discovery (repo: https://github.com/qhjqhj00/MemoRAG, paper: https://arxiv.org/pdf/2409.05591v1) :fire:
 - 9/2/2024: Start to maintain the [tutorials](./Tutorials/). The contents within will be actively updated and eariched, stay tuned! :books:
 - 7/26/2024: Release a new embedding model [bge-en-icl](https://huggingface.co/BAAI/bge-en-icl), an embedding model that incorporates in-context learning capabilities, which, by providing task-relevant query-response examples, can encode semantically richer queries, further enhancing the semantic representation ability of the embeddings. :fire:
 - 7/26/2024: Release a new embedding model [bge-multilingual-gemma2](https://huggingface.co/BAAI/bge-multilingual-gemma2), a multilingual embedding model based on gemma-2-9b, which supports multiple languages and diverse downstream tasks, achieving new SOTA on multilingual benchmarks (MIRACL, MTEB-fr, and MTEB-pl). :fire:

Tutorials/2_Similarity/similarity.ipynb

@@ -187,7 +187,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "tensor([[3., 2., 3., 6.]]) tensor([[2., 4., 4., 1.]])\n"
+      "tensor([[5., 2., 2., 6.]]) tensor([[4., 6., 6., 4.]])\n"
      ]
     }
    ],
@@ -239,7 +239,7 @@
     {
      "data": {
       "text/plain": [
-       "5.5677642822265625"
+       "6.082762718200684"
       ]
      },
      "execution_count": 6,
@@ -273,7 +273,7 @@
     {
      "data": {
       "text/plain": [
-       "5.5677642822265625"
+       "6.082762718200684"
       ]
      },
      "execution_count": 7,
@@ -382,7 +382,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "0.6907725930213928\n"
+      "0.802726686000824\n"
      ]
     }
    ],
@@ -439,7 +439,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "0.6907725930213928\n"
+      "0.802726686000824\n"
      ]
     }
    ],
@@ -487,7 +487,7 @@
     {
      "data": {
       "text/plain": [
-       "0.6907725930213928"
+       "0.802726686000824"
       ]
      },
      "execution_count": 11,
@@ -530,7 +530,7 @@
     {
      "data": {
       "text/plain": [
-       "32.0"
+       "68.0"
       ]
      },
      "execution_count": 12,
@@ -563,31 +563,31 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 20,
+   "execution_count": 14,
    "id": "e0f40534",
    "metadata": {},
    "outputs": [],
    "source": [
     "from FlagEmbedding import FlagModel\n",
     "\n",
-    "model = FlagModel('BAAI/bge-base-en-v1.5',\n",
+    "model = FlagModel('BAAI/bge-large-en-v1.5',\n",
     "                  query_instruction_for_retrieval=\"Represent this sentence for searching relevant passages:\",\n",
     "                  use_fp16=True)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 21,
+   "execution_count": 15,
    "id": "78445a86",
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "0.9999999403953552"
+       "1.0"
       ]
      },
-     "execution_count": 21,
+     "execution_count": 15,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -616,7 +616,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 22,
+   "execution_count": 16,
    "id": "73012cbb",
    "metadata": {},
    "outputs": [],
@@ -648,7 +648,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 23,
+   "execution_count": 17,
    "id": "98bfcc6d",
    "metadata": {},
    "outputs": [
@@ -684,15 +684,15 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 24,
+   "execution_count": 18,
    "id": "426c0b42",
    "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "torch.Size([1, 768])\n"
+      "torch.Size([1, 1024])\n"
      ]
     }
    ],
@@ -715,16 +715,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 25,
+   "execution_count": 19,
    "id": "d9bb35cf",
    "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "0.49946388602256775\n",
-      "0.6032702922821045\n"
+      "0.714613139629364\n",
+      "0.5931472182273865\n"
      ]
     }
    ],
@@ -745,16 +745,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 26,
+   "execution_count": 20,
    "id": "29e70bbc",
    "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "0.8752679824829102\n",
-      "0.8180326223373413\n"
+      "0.7446640729904175\n",
+      "0.8240882158279419\n"
      ]
     }
    ],

Tutorials/3_Indexing/indexing.ipynb → Tutorials/3_Indexing/3.1.1_Intro_to_Faiss.ipynb

@@ -4,7 +4,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Indexing"
+    "# Indexing Using Faiss"
    ]
   },
   {