From 96843e5285aa6dc65c1c0efb8e56e809f129d772 Mon Sep 17 00:00:00 2001 From: Samuel Yang Date: Tue, 7 Jan 2025 13:56:29 +0800 Subject: [PATCH 1/5] feat(models): add infly/inf-retriever-v1 model metadata- Add inf_models.py file with metadata for infly/inf-retriever-v1 model - Update overview.py to include inf_models in model imports --- mteb/models/inf_models.py | 28 ++++++++++++++++++++++++++++ mteb/models/overview.py | 2 ++ 2 files changed, 30 insertions(+) create mode 100644 mteb/models/inf_models.py diff --git a/mteb/models/inf_models.py b/mteb/models/inf_models.py new file mode 100644 index 0000000000..58c4c93f78 --- /dev/null +++ b/mteb/models/inf_models.py @@ -0,0 +1,28 @@ +from __future__ import annotations + +from functools import partial + +from mteb.model_meta import ModelMeta, sentence_transformers_loader + +inf_retriever_v1 = ModelMeta( + loader=partial( # type: ignore + sentence_transformers_loader, + model_name="infly/inf-retriever-v1", + revision="d2d074546028c0012b5cc6af78c4fac24896e67f", + trust_remote_code=True, + ), + name="infly/inf-retriever-v1", + languages=["eng_Latn","zho_Hans"], + open_weights=True, + revision="d2d074546028c0012b5cc6af78c4fac24896e67f", + release_date="2024-12-24", # initial commit of hf model. + n_parameters=7_069_121_024, + memory_usage=None, + embed_dim=3584, + license=None, + max_tokens=131_072, + reference="https://huggingface.co/infly/inf-retriever-v1", + similarity_fn_name="cosine", + framework=["Sentence Transformers", "PyTorch"], + use_instructions=True, +) \ No newline at end of file diff --git a/mteb/models/overview.py b/mteb/models/overview.py index 5a3cfe757f..d9a2041601 100644 --- a/mteb/models/overview.py +++ b/mteb/models/overview.py @@ -22,6 +22,7 @@ gritlm_models, gte_models, ibm_granite_models, + inf_models, jasper_models, jina_models, linq_models, @@ -60,6 +61,7 @@ gritlm_models, gte_models, ibm_granite_models, + inf_models, jina_models, linq_models, llm2vec_models, From f3ce39f69bdf8285dd02ff724e35f591de20aa35 Mon Sep 17 00:00:00 2001 From: Samuel Yang Date: Thu, 9 Jan 2025 15:55:14 +0800 Subject: [PATCH 2/5] Reformat code --- mteb/models/inf_models.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/mteb/models/inf_models.py b/mteb/models/inf_models.py index 58c4c93f78..b9dc4495de 100644 --- a/mteb/models/inf_models.py +++ b/mteb/models/inf_models.py @@ -12,7 +12,7 @@ trust_remote_code=True, ), name="infly/inf-retriever-v1", - languages=["eng_Latn","zho_Hans"], + languages=["eng_Latn", "zho_Hans"], open_weights=True, revision="d2d074546028c0012b5cc6af78c4fac24896e67f", release_date="2024-12-24", # initial commit of hf model. @@ -25,4 +25,4 @@ similarity_fn_name="cosine", framework=["Sentence Transformers", "PyTorch"], use_instructions=True, -) \ No newline at end of file +) From 496901fc482768e0c8c2c92e5d7f63512f14a1eb Mon Sep 17 00:00:00 2001 From: Samuel Yang Date: Fri, 10 Jan 2025 14:12:25 +0800 Subject: [PATCH 3/5] Update inf-retriever-v1 ModelMeta --- mteb/models/inf_models.py | 1 + 1 file changed, 1 insertion(+) diff --git a/mteb/models/inf_models.py b/mteb/models/inf_models.py index b9dc4495de..c48d1e69a4 100644 --- a/mteb/models/inf_models.py +++ b/mteb/models/inf_models.py @@ -25,4 +25,5 @@ similarity_fn_name="cosine", framework=["Sentence Transformers", "PyTorch"], use_instructions=True, + adapted_from="Alibaba-NLP/gte-Qwen2-7B-instruct", ) From 00c814f0d8b532d5d0ff7b50facefe72c20d052e Mon Sep 17 00:00:00 2001 From: Samuel Yang Date: Mon, 13 Jan 2025 13:56:13 +0800 Subject: [PATCH 4/5] Fill more information for inf-retriever-v1 --- mteb/models/inf_models.py | 3 +++ 1 file changed, 3 insertions(+) diff --git a/mteb/models/inf_models.py b/mteb/models/inf_models.py index c48d1e69a4..6e66b4c5f1 100644 --- a/mteb/models/inf_models.py +++ b/mteb/models/inf_models.py @@ -26,4 +26,7 @@ framework=["Sentence Transformers", "PyTorch"], use_instructions=True, adapted_from="Alibaba-NLP/gte-Qwen2-7B-instruct", + public_training_code=False, + public_training_data=False, + training_datasets=None, ) From d689be50f285039b84ec7ddc24375730021b45b6 Mon Sep 17 00:00:00 2001 From: Samuel Yang Date: Wed, 15 Jan 2025 14:38:42 +0800 Subject: [PATCH 5/5] Add license information for inf-retriever-v1 --- mteb/models/inf_models.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mteb/models/inf_models.py b/mteb/models/inf_models.py index 6e66b4c5f1..4670b20735 100644 --- a/mteb/models/inf_models.py +++ b/mteb/models/inf_models.py @@ -19,7 +19,7 @@ n_parameters=7_069_121_024, memory_usage=None, embed_dim=3584, - license=None, + license="apache-2.0", max_tokens=131_072, reference="https://huggingface.co/infly/inf-retriever-v1", similarity_fn_name="cosine",