Upload folder using huggingface_hub

.ipynb_checkpoints/README-checkpoint.md

@@ -0,0 +1,107 @@
+---
+language:
+- en
+- sv
+tags:
+- translation
+license: cc-by-4.0
+datasets:
+- quickmt/quickmt-train.sv-en
+model-index:
+- name: quickmt-sv-en
+  results:
+  - task:
+      name: Translation swe-eng
+      type: translation
+      args: swe-eng
+    dataset:
+      name: flores101-devtest
+      type: flores_101
+      args: swe_Latn eng_Latn devtest
+    metrics:
+    - name: BLEU
+      type: bleu
+      value: 47.59
+    - name: CHRF
+      type: chrf
+      value: 47.59
+    - name: COMET
+      type: comet
+      value: 47.59
+---
+
+
+# `quickmt-sv-en` Neural Machine Translation Model 
+
+`quickmt-sv-en` is a reasonably fast and reasonably accurate neural machine translation model for translation from `sv` into `en`.
+
+
+## Try it on our Huggingface Space
+
+Give it a try before downloading here: https://huggingface.co/spaces/quickmt/QuickMT-Demo
+
+
+## Model Information
+
+* Trained using [`eole`](https://github.com/eole-nlp/eole) 
+* 200M parameter transformer 'big' with 8 encoder layers and 2 decoder layers
+* 32k separate Sentencepiece vocabs
+* Exported for fast inference to [CTranslate2](https://github.com/OpenNMT/CTranslate2) format
+* The pytorch model (for use with [`eole`](https://github.com/eole-nlp/eole)) is available in this repository in the `eole-model` folder
+
+See the `eole` model configuration in this repository for further details and the `eole-model` for the raw `eole` (pytorch) model.
+
+
+## Usage with `quickmt`
+
+You must install the Nvidia cuda toolkit first, if you want to do GPU inference.
+
+Next, install the `quickmt` python library and download the model:
+
+```bash
+git clone https://github.com/quickmt/quickmt.git
+pip install ./quickmt/
+
+quickmt-model-download quickmt/quickmt-sv-en ./quickmt-sv-en
+```
+
+Finally use the model in python:
+
+```python
+from quickmt import Translator
+
+# Auto-detects GPU, set to "cpu" to force CPU inference
+t = Translator("./quickmt-sv-en/", device="auto")
+
+# Translate - set beam size to 1 for faster speed (but lower quality)
+sample_text = 'Dr. Ehud Ur, professor i medicin vid Dalhousie University i Halifax, Nova Scotia och ordförande för den kliniska och vetenskapliga avdelningen av  den Kanadensiska diabetesföreningen, varnade för att forskningen fortfarande befinner sig i ett tidigt stadium.'
+
+t(sample_text, beam_size=5)
+```
+
+> 'Dr. Ehud Ur, professor of medicine at Dalhousie University in Halifax, Nova Scotia and chairman of the clinical and scientific department of the Canadian Diabetes Association, warned that the research is still at an early stage.'
+
+```python
+# Get alternative translations by sampling
+# You can pass any cTranslate2 `translate_batch` arguments
+t([sample_text], sampling_temperature=1.2, beam_size=1, sampling_topk=50, sampling_topp=0.9)
+```
+
+> 'Dr. Ehud Ur, a Professor of Medicine at Dalhousie University in Halifax, Nova Scotia and Chair of the Clinical and Scientific Division of the Canadian Diabetes Society, warned that the research is still at an early stage.'
+
+The model is in `ctranslate2` format, and the tokenizers are `sentencepiece`, so you can use `ctranslate2` directly instead of through `quickmt`. It is also possible  to get this model to work with e.g. [LibreTranslate](https://libretranslate.com/) which also uses `ctranslate2` and `sentencepiece`. A model in safetensors format to be used with `eole` is also provided.
+
+
+## Metrics
+
+`bleu` and `chrf2` are calculated with [sacrebleu](https://github.com/mjpost/sacrebleu) on the [Flores200 `devtest` test set](https://huggingface.co/datasets/facebook/flores) ("swe_Latn"->"eng_Latn"). `comet22` with the [`comet`](https://github.com/Unbabel/COMET) library and the [default model](https://huggingface.co/Unbabel/wmt22-comet-da). "Time (s)" is the time in seconds to translate the flores-devtest dataset (1012 sentences) on an RTX 4070s GPU with batch size 32.
+
+
+|                                  |   bleu |   chrf2 |   comet22 |   Time (s) |
+|:---------------------------------|-------:|--------:|----------:|-----------:|
+| quickmt/quickmt-sv-en            |  47.59 |   70.93 |     89.82 |       1.5  |
+| Helsinki-NLP/opus-mt-sv-en       |  45.51 |   68.88 |     89.08 |       3.25 |
+| facebook/nllb-200-distilled-600M |  46.69 |   69.22 |     89.17 |      20.82 |
+| facebook/nllb-200-distilled-1.3B |  49.29 |   71.12 |     89.99 |      36.76 |
+| facebook/m2m100_418M             |  40.05 |   65.13 |     85.91 |      17.6  |
+| facebook/m2m100_1.2B             |  45.34 |   68.78 |     88.95 |      34.15 |

.ipynb_checkpoints/eole-config-checkpoint.yaml

@@ -0,0 +1,95 @@
+## IO
+save_data: data
+overwrite: True
+seed: 1234
+report_every: 100
+valid_metrics: ["BLEU"]
+tensorboard: true
+tensorboard_log_dir: tensorboard
+
+### Vocab
+src_vocab: sv.eole.vocab
+tgt_vocab: en.eole.vocab
+src_vocab_size: 32000
+tgt_vocab_size: 32000
+vocab_size_multiple: 8
+share_vocab: false
+n_sample: 0
+
+data:
+    corpus_1:
+        path_src: hf://quickmt/quickmt-train.sv-en/sv
+        path_tgt: hf://quickmt/quickmt-train.sv-en/en
+        path_sco: hf://quickmt/quickmt-train.sv-en/sco
+    valid:
+        path_src: valid.sv
+        path_tgt: valid.en
+
+transforms: [sentencepiece, filtertoolong]
+transforms_configs:
+  sentencepiece:
+    src_subword_model: "sv.spm.model"
+    tgt_subword_model: "en.spm.model"
+  filtertoolong:
+    src_seq_length: 256
+    tgt_seq_length: 256
+
+training:
+    # Run configuration
+    model_path: quickmt-sv-en-eole-model
+    keep_checkpoint: 4
+    train_steps: 100000
+    save_checkpoint_steps: 5000
+    valid_steps: 5000
+    
+    # Train on a single GPU
+    world_size: 1
+    gpu_ranks: [0]
+
+    # Batching 10240
+    batch_type: "tokens"
+    batch_size: 6000
+    valid_batch_size: 2048
+    batch_size_multiple: 8
+    accum_count: [20]
+    accum_steps: [0]
+
+    # Optimizer & Compute
+    compute_dtype: "fp16"
+    optim: "adamw"
+    #use_amp: False
+    learning_rate: 2.0
+    warmup_steps: 2000
+    decay_method: "noam"
+    adam_beta2: 0.998
+
+    # Data loading
+    bucket_size: 128000
+    num_workers: 4
+    prefetch_factor: 32
+    
+    # Hyperparams
+    dropout_steps: [0]
+    dropout: [0.1]
+    attention_dropout: [0.1]
+    max_grad_norm: 0
+    label_smoothing: 0.1
+    average_decay: 0.0001
+    param_init_method: xavier_uniform
+    normalization: "tokens"
+
+model:
+    architecture: "transformer"
+    share_embeddings: false
+    share_decoder_embeddings: true
+    hidden_size: 1024
+    encoder:
+        layers: 8
+    decoder:
+        layers: 2
+    heads: 8
+    transformer_ff: 4096
+    embeddings:
+        word_vec_size: 1024
+        position_encoding_type: "SinusoidalInterleaved"
+

README.md

@@ -1,3 +1,107 @@
----
-license: cc-by-4.0
----
+---
+language:
+- en
+- sv
+tags:
+- translation
+license: cc-by-4.0
+datasets:
+- quickmt/quickmt-train.sv-en
+model-index:
+- name: quickmt-sv-en
+  results:
+  - task:
+      name: Translation swe-eng
+      type: translation
+      args: swe-eng
+    dataset:
+      name: flores101-devtest
+      type: flores_101
+      args: swe_Latn eng_Latn devtest
+    metrics:
+    - name: BLEU
+      type: bleu
+      value: 47.59
+    - name: CHRF
+      type: chrf
+      value: 47.59
+    - name: COMET
+      type: comet
+      value: 47.59
+---
+
+
+# `quickmt-sv-en` Neural Machine Translation Model 
+
+`quickmt-sv-en` is a reasonably fast and reasonably accurate neural machine translation model for translation from `sv` into `en`.
+
+
+## Try it on our Huggingface Space
+
+Give it a try before downloading here: https://huggingface.co/spaces/quickmt/QuickMT-Demo
+
+
+## Model Information
+
+* Trained using [`eole`](https://github.com/eole-nlp/eole) 
+* 200M parameter transformer 'big' with 8 encoder layers and 2 decoder layers
+* 32k separate Sentencepiece vocabs
+* Exported for fast inference to [CTranslate2](https://github.com/OpenNMT/CTranslate2) format
+* The pytorch model (for use with [`eole`](https://github.com/eole-nlp/eole)) is available in this repository in the `eole-model` folder
+
+See the `eole` model configuration in this repository for further details and the `eole-model` for the raw `eole` (pytorch) model.
+
+
+## Usage with `quickmt`
+
+You must install the Nvidia cuda toolkit first, if you want to do GPU inference.
+
+Next, install the `quickmt` python library and download the model:
+
+```bash
+git clone https://github.com/quickmt/quickmt.git
+pip install ./quickmt/
+
+quickmt-model-download quickmt/quickmt-sv-en ./quickmt-sv-en
+```
+
+Finally use the model in python:
+
+```python
+from quickmt import Translator
+
+# Auto-detects GPU, set to "cpu" to force CPU inference
+t = Translator("./quickmt-sv-en/", device="auto")
+
+# Translate - set beam size to 1 for faster speed (but lower quality)
+sample_text = 'Dr. Ehud Ur, professor i medicin vid Dalhousie University i Halifax, Nova Scotia och ordförande för den kliniska och vetenskapliga avdelningen av  den Kanadensiska diabetesföreningen, varnade för att forskningen fortfarande befinner sig i ett tidigt stadium.'
+
+t(sample_text, beam_size=5)
+```
+
+> 'Dr. Ehud Ur, professor of medicine at Dalhousie University in Halifax, Nova Scotia and chairman of the clinical and scientific department of the Canadian Diabetes Association, warned that the research is still at an early stage.'
+
+```python
+# Get alternative translations by sampling
+# You can pass any cTranslate2 `translate_batch` arguments
+t([sample_text], sampling_temperature=1.2, beam_size=1, sampling_topk=50, sampling_topp=0.9)
+```
+
+> 'Dr. Ehud Ur, a Professor of Medicine at Dalhousie University in Halifax, Nova Scotia and Chair of the Clinical and Scientific Division of the Canadian Diabetes Society, warned that the research is still at an early stage.'
+
+The model is in `ctranslate2` format, and the tokenizers are `sentencepiece`, so you can use `ctranslate2` directly instead of through `quickmt`. It is also possible  to get this model to work with e.g. [LibreTranslate](https://libretranslate.com/) which also uses `ctranslate2` and `sentencepiece`. A model in safetensors format to be used with `eole` is also provided.
+
+
+## Metrics
+
+`bleu` and `chrf2` are calculated with [sacrebleu](https://github.com/mjpost/sacrebleu) on the [Flores200 `devtest` test set](https://huggingface.co/datasets/facebook/flores) ("swe_Latn"->"eng_Latn"). `comet22` with the [`comet`](https://github.com/Unbabel/COMET) library and the [default model](https://huggingface.co/Unbabel/wmt22-comet-da). "Time (s)" is the time in seconds to translate the flores-devtest dataset (1012 sentences) on an RTX 4070s GPU with batch size 32.
+
+
+|                                  |   bleu |   chrf2 |   comet22 |   Time (s) |
+|:---------------------------------|-------:|--------:|----------:|-----------:|
+| quickmt/quickmt-sv-en            |  47.59 |   70.93 |     89.82 |       1.5  |
+| Helsinki-NLP/opus-mt-sv-en       |  45.51 |   68.88 |     89.08 |       3.25 |
+| facebook/nllb-200-distilled-600M |  46.69 |   69.22 |     89.17 |      20.82 |
+| facebook/nllb-200-distilled-1.3B |  49.29 |   71.12 |     89.99 |      36.76 |
+| facebook/m2m100_418M             |  40.05 |   65.13 |     85.91 |      17.6  |
+| facebook/m2m100_1.2B             |  45.34 |   68.78 |     88.95 |      34.15 |

config.json

@@ -0,0 +1,10 @@
+{
+  "add_source_bos": false,
+  "add_source_eos": false,
+  "bos_token": "<s>",
+  "decoder_start_token": "<s>",
+  "eos_token": "</s>",
+  "layer_norm_epsilon": 1e-06,
+  "multi_query_attention": false,
+  "unk_token": "<unk>"
+}

eole-config.yaml

@@ -0,0 +1,95 @@
+## IO
+save_data: data
+overwrite: True
+seed: 1234
+report_every: 100
+valid_metrics: ["BLEU"]
+tensorboard: true
+tensorboard_log_dir: tensorboard
+
+### Vocab
+src_vocab: sv.eole.vocab
+tgt_vocab: en.eole.vocab
+src_vocab_size: 32000
+tgt_vocab_size: 32000
+vocab_size_multiple: 8
+share_vocab: false
+n_sample: 0
+
+data:
+    corpus_1:
+        path_src: hf://quickmt/quickmt-train.sv-en/sv
+        path_tgt: hf://quickmt/quickmt-train.sv-en/en
+        path_sco: hf://quickmt/quickmt-train.sv-en/sco
+    valid:
+        path_src: valid.sv
+        path_tgt: valid.en
+
+transforms: [sentencepiece, filtertoolong]
+transforms_configs:
+  sentencepiece:
+    src_subword_model: "sv.spm.model"
+    tgt_subword_model: "en.spm.model"
+  filtertoolong:
+    src_seq_length: 256
+    tgt_seq_length: 256
+
+training:
+    # Run configuration
+    model_path: quickmt-sv-en-eole-model
+    keep_checkpoint: 4
+    train_steps: 100000
+    save_checkpoint_steps: 5000
+    valid_steps: 5000
+    
+    # Train on a single GPU
+    world_size: 1
+    gpu_ranks: [0]
+
+    # Batching 10240
+    batch_type: "tokens"
+    batch_size: 6000
+    valid_batch_size: 2048
+    batch_size_multiple: 8
+    accum_count: [20]
+    accum_steps: [0]
+
+    # Optimizer & Compute
+    compute_dtype: "fp16"
+    optim: "adamw"
+    #use_amp: False
+    learning_rate: 2.0
+    warmup_steps: 2000
+    decay_method: "noam"
+    adam_beta2: 0.998
+
+    # Data loading
+    bucket_size: 128000
+    num_workers: 4
+    prefetch_factor: 32
+    
+    # Hyperparams
+    dropout_steps: [0]
+    dropout: [0.1]
+    attention_dropout: [0.1]
+    max_grad_norm: 0
+    label_smoothing: 0.1
+    average_decay: 0.0001
+    param_init_method: xavier_uniform
+    normalization: "tokens"
+
+model:
+    architecture: "transformer"
+    share_embeddings: false
+    share_decoder_embeddings: true
+    hidden_size: 1024
+    encoder:
+        layers: 8
+    decoder:
+        layers: 2
+    heads: 8
+    transformer_ff: 4096
+    embeddings:
+        word_vec_size: 1024
+        position_encoding_type: "SinusoidalInterleaved"
+

eole-model/config.json

@@ -0,0 +1,132 @@
+{
+  "tgt_vocab_size": 32000,
+  "n_sample": 0,
+  "tgt_vocab": "en.eole.vocab",
+  "overwrite": true,
+  "src_vocab": "sv.eole.vocab",
+  "src_vocab_size": 32000,
+  "save_data": "data",
+  "tensorboard_log_dir_dated": "tensorboard/Nov-09_21-14-34",
+  "report_every": 100,
+  "share_vocab": false,
+  "tensorboard": true,
+  "transforms": [
+    "sentencepiece",
+    "filtertoolong"
+  ],
+  "seed": 1234,
+  "valid_metrics": [
+    "BLEU"
+  ],
+  "tensorboard_log_dir": "tensorboard",
+  "vocab_size_multiple": 8,
+  "training": {
+    "optim": "adamw",
+    "world_size": 1,
+    "batch_size": 6000,
+    "param_init_method": "xavier_uniform",
+    "batch_size_multiple": 8,
+    "label_smoothing": 0.1,
+    "dropout_steps": [
+      0
+    ],
+    "bucket_size": 128000,
+    "adam_beta2": 0.998,
+    "compute_dtype": "torch.float16",
+    "dropout": [
+      0.1
+    ],
+    "valid_batch_size": 2048,
+    "model_path": "quickmt-sv-en-eole-model",
+    "valid_steps": 5000,
+    "average_decay": 0.0001,
+    "decay_method": "noam",
+    "batch_type": "tokens",
+    "prefetch_factor": 32,
+    "train_steps": 100000,
+    "num_workers": 0,
+    "normalization": "tokens",
+    "attention_dropout": [
+      0.1
+    ],
+    "warmup_steps": 2000,
+    "accum_steps": [
+      0
+    ],
+    "accum_count": [
+      20
+    ],
+    "max_grad_norm": 0.0,
+    "save_checkpoint_steps": 5000,
+    "keep_checkpoint": 4,
+    "learning_rate": 2.0,
+    "gpu_ranks": [
+      0
+    ]
+  },
+  "model": {
+    "architecture": "transformer",
+    "share_decoder_embeddings": true,
+    "hidden_size": 1024,
+    "share_embeddings": false,
+    "heads": 8,
+    "transformer_ff": 4096,
+    "position_encoding_type": "SinusoidalInterleaved",
+    "encoder": {
+      "n_positions": null,
+      "hidden_size": 1024,
+      "layers": 8,
+      "heads": 8,
+      "encoder_type": "transformer",
+      "transformer_ff": 4096,
+      "src_word_vec_size": 1024,
+      "position_encoding_type": "SinusoidalInterleaved"
+    },
+    "decoder": {
+      "tgt_word_vec_size": 1024,
+      "n_positions": null,
+      "hidden_size": 1024,
+      "layers": 2,
+      "heads": 8,
+      "decoder_type": "transformer",
+      "transformer_ff": 4096,
+      "position_encoding_type": "SinusoidalInterleaved"
+    },
+    "embeddings": {
+      "position_encoding_type": "SinusoidalInterleaved",
+      "tgt_word_vec_size": 1024,
+      "src_word_vec_size": 1024,
+      "word_vec_size": 1024
+    }
+  },
+  "data": {
+    "corpus_1": {
+      "transforms": [
+        "sentencepiece",
+        "filtertoolong"
+      ],
+      "path_src": "train.sv",
+      "path_tgt": "train.en",
+      "path_align": null
+    },
+    "valid": {
+      "transforms": [
+        "sentencepiece",
+        "filtertoolong"
+      ],
+      "path_src": "valid.sv",
+      "path_tgt": "valid.en",
+      "path_align": null
+    }
+  },
+  "transforms_configs": {
+    "sentencepiece": {
+      "src_subword_model": "${MODEL_PATH}/sv.spm.model",
+      "tgt_subword_model": "${MODEL_PATH}/en.spm.model"
+    },
+    "filtertoolong": {
+      "src_seq_length": 256,
+      "tgt_seq_length": 256
+    }
+  }
+}

eole-model/en.spm.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0eeaccede2c05786b37d496470fbfc1e0509bf61be3e16913981d3a195873bdf
+size 800835

eole-model/model.00.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbc8dc4e582896516e3bc64dac2d34b0870e3afa9dfc78321377a1574bc0986e
+size 840314816

eole-model/sv.spm.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de8b005b6a57ec60d00845c007527d6e6d2bcedaabcc842f15e579b294e5250c
+size 814642

model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b6a7400b4c0ece91190c8ae780f31f208b8d33bf469dfd9dcb06b5323220c10
+size 409915789

src.spm.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de8b005b6a57ec60d00845c007527d6e6d2bcedaabcc842f15e579b294e5250c
+size 814642

tgt.spm.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0eeaccede2c05786b37d496470fbfc1e0509bf61be3e16913981d3a195873bdf
+size 800835