push model

.gitattributes

@@ -33,3 +33,4 @@ 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
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,3 +1,97 @@
 ---
+language: 
+- en
+- de
+- es
+- fr
+tags:
+- clir
+- colbertx
+- plaidx
+- xlm-roberta-large
+datasets:
+- ms_marco
+- hltcoe/tdist-msmarco-scores
+task_categories:
+- text-retrieval
+- information-retrieval
+task_ids:
+- passage-retrieval
+- cross-language-retrieval
 license: mit
 ---
+
+# ColBERT-X for English-German/Spanish/French MLIR using Multilingual Translate-Distill
+
+## CLIR Model Setting
+
+- Query language: English
+- Query length: 32 token max
+- Document language: German/Spanish/French
+- Document length: 180 token max (please use MaxP to aggregate the passage score if needed)
+
+## Model Description
+
+Multilingual Translate-Distill is a training technique that produces state-of-the-art MLIR dense retrieval model through translation and distillation.
+`plaidx-large-clef-mtd-mix-passages-mt5xxl-engeng` is trained with KL-Divergence from the `mt5xxl` MonoT5 reranker
+[`unicamp-dl/mt5-13b-mmarco-100k`](https://huggingface.co/unicamp-dl/mt5-13b-mmarco-100k)
+inferenced on English MS MARCO training queries and passages. 
+The teacher scores can be found in 
+[`hltcoe/tdist-msmarco-scores`](https://huggingface.co/datasets/hltcoe/tdist-msmarco-scores/blob/main/t53b-monot5-msmarco-engeng.jsonl.gz). 
+
+### Training Parameters
+
+- learning rate: 5e-6
+- update steps: 200,000
+- nway (number of passages per query): 6 (randomly selected from 50; 2 if using `round-robin-entires`, see below)
+- per device batch size (number of query-passage set): 8
+- training GPU: 8 NVIDIA V100 with 32 GB memory
+
+### Mixing Strategies
+
+- `mix-passages`: languages are randomly assigned to the 6 sampled passages for a given query during training. 
+- `mix-entries`: all passages in the a given query-passage set are randomly assigned to the same language. 
+- `round-robin-entires`: for each query, the query-passage set is repeated `n` times to iterate through all languages. 
+
+## Usage
+
+To properly load ColBERT-X models from Huggingface Hub, please use the following version of PLAID-X. 
+```bash
+pip install PLAID-X>=0.3.1
+```
+
+Following code snippet loads the model through Huggingface API. 
+```python
+from colbert.modeling.checkpoint import Checkpoint
+from colbert.infra import ColBERTConfig
+
+Checkpoint('hltcoe/plaidx-large-clef-mtd-mix-passages-mt5xxl-engeng', colbert_config=ColBERTConfig())
+```
+
+For full tutorial, please refer to the [PLAID-X Jupyter Notebook](https://colab.research.google.com/github/hltcoe/clir-tutorial/blob/main/notebooks/clir_tutorial_plaidx.ipynb), 
+which is part of the [SIGIR 2023 CLIR Tutorial](https://github.com/hltcoe/clir-tutorial). 
+
+## BibTeX entry and Citation Info
+
+Please cite the following two papers if you use the model. 
+
+
+```bibtex
+@inproceedings{mtt,
+	title = {Neural Approaches to Multilingual Information Retrieval},
+	author = {Dawn Lawrie and Eugene Yang and Douglas W Oard and James Mayfield},
+	booktitle = {Proceedings of the 45th European Conference on Information Retrieval (ECIR)},
+	year = {2023},
+    doi = {10.1007/978-3-031-28244-7_33},
+	url = {https://arxiv.org/abs/2209.01335}
+}
+```
+
+```bibtex
+@inproceedings{mtd,
+	author = {Eugene Yang and Dawn Lawrie and James Mayfield},
+	title = {Distillation for Multilingual Information Retrieval},
+	booktitle = {Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) (Short Paper) (Accepted)},
+	year = {2024}
+}
+```

added_tokens.json

@@ -0,0 +1,4 @@
+{
+  "[unused0]": 250002,
+  "[unused1]": 250003
+}

artifact.metadata

@@ -0,0 +1,76 @@
+{
+    "query_token_id": "[unused0]",
+    "doc_token_id": "[unused1]",
+    "query_token": "[Q]",
+    "doc_token": "[D]",
+    "ncells": null,
+    "centroid_score_threshold": null,
+    "ndocs": null,
+    "only_approx": false,
+    "index_path": null,
+    "nbits": 1,
+    "kmeans_niters": 4,
+    "resume": false,
+    "max_sampled_pid": -1,
+    "max_num_partitions": -1,
+    "use_lagacy_build_ivf": false,
+    "reuse_centroids_from": null,
+    "similarity": "cosine",
+    "bsize": 8,
+    "accumsteps": 1,
+    "lr": 5e-6,
+    "maxsteps": 400000,
+    "save_every": null,
+    "resume_optimizer": false,
+    "fix_broken_optimizer_state": false,
+    "warmup": null,
+    "warmup_bert": null,
+    "relu": false,
+    "nway": 6,
+    "n_query_alternative": 1,
+    "use_ib_negatives": false,
+    "kd_loss": "KLD",
+    "reranker": false,
+    "distillation_alpha": 1.0,
+    "ignore_scores": false,
+    "model_name": "xlm-roberta-large",
+    "force_resize_embeddings": true,
+    "shuffle_passages": true,
+    "sampling_max_beta": 1.0,
+    "over_one_epoch": true,
+    "multilang": false,
+    "nolangreg": true,
+    "query_maxlen": 32,
+    "attend_to_mask_tokens": false,
+    "interaction": "colbert",
+    "dim": 128,
+    "doc_maxlen": 220,
+    "mask_punctuation": true,
+    "checkpoint": "xlm-roberta-large",
+    "triples": "\/expscratch\/eyang\/workspace\/plaid-aux\/training_triples\/msmarco-passages\/triples_mt5xxl-monot5-mmarco-engeng.jsonl",
+    "collection": "Combination(elements)[.\/mt_msmarco\/de\/collection_passages.tsv+.\/mt_msmarco\/es\/collection_passages.tsv+.\/mt_msmarco\/fr\/collection_passages.tsv]",
+    "queries": "irds:msmarco-passage\/train:queries",
+    "index_name": null,
+    "debug": false,
+    "overwrite": false,
+    "root": "\/expscratch\/eyang\/workspace\/plaid-aux\/experiments",
+    "experiment": "mtt-tdistill",
+    "index_root": null,
+    "name": "multi-KLD-shuf-5e-6\/mt5xxl-monot5-mmarco-engeng.de+es+fr\/64bat.6way",
+    "rank": 0,
+    "nranks": 8,
+    "amp": true,
+    "ivf_num_processes": 20,
+    "ivf_use_tempdir": false,
+    "ivf_merging_ways": 2,
+    "gpus": 8,
+    "meta": {
+        "hostname": "r8n06",
+        "git_branch": "eugene-training",
+        "git_hash": "683c265c010eb72e00617bdb6771b333acfa0191",
+        "git_commit_datetime": "2024-01-04 14:11:31-05:00",
+        "current_datetime": "Jan 19, 2024 ;  2:10AM EST (-0500)",
+        "cmd": "train.py --model_name xlm-roberta-large --training_triples \/expscratch\/eyang\/workspace\/plaid-aux\/training_triples\/msmarco-passages\/triples_mt5xxl-monot5-mmarco-engeng.jsonl --training_queries msmarco-passage\/train --training_collection .\/mt_msmarco\/de\/collection_passages.tsv .\/mt_msmarco\/es\/collection_passages.tsv .\/mt_msmarco\/fr\/collection_passages.tsv --training_collection_mixing elements --other_args nolangreg=True --maxsteps 400000 --learning_rate 5e-6 --kd_loss KLD --per_device_batch_size 8 --nway 6 --run_tag multi-KLD-shuf-5e-6\/mt5xxl-monot5-mmarco-engeng.de+es+fr --experiment mtt-tdistill",
+        "version": "colbert-v0.4"
+    }
+}

config.json

@@ -0,0 +1,28 @@
+{
+  "_name_or_path": "xlm-roberta-large",
+  "architectures": [
+    "HF_ColBERT"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "xlm-roberta",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "output_past": true,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.28.0",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 250004
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2287cdbb3de08b977a27a0cd8d37181be3d458e0c3f63aeb00df7e3b2ea0fd66
+size 2240233969

sentencepiece.bpe.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
+size 5069051

special_tokens_map.json

@@ -0,0 +1,15 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c19ec03247ee31e5f42772ac32bde8dca2727b30c8310c2e585df4980a8db230
+size 17083032

tokenizer_config.json

@@ -0,0 +1,19 @@
+{
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "__type": "AddedToken",
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "model_max_length": 512,
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "tokenizer_class": "XLMRobertaTokenizer",
+  "unk_token": "<unk>"
+}