Add ctransformers library

docs/hub/_toctree.yml

@@ -59,6 +59,8 @@
         title: AllenNLP
       - local: asteroid
         title: Asteroid
+      - local: ctransformers
+        title: CTransformers
       - local: espnet
         title: ESPnet
       - local: fastai

docs/hub/ctransformers.md

@@ -0,0 +1,40 @@
+# Using `ctransformers` at Hugging Face
+
+`ctransformers` is a library that provides Python bindings for the Transformer models implemented in C/C++ using GGML library. It supports several state-of-the-art language [models](https://github.com/marella/ctransformers#supported-models).
+
+## Exploring `ctransformers` in the Hub
+
+You can find `ctransformers` models by filtering at the left of the [models page](https://huggingface.co/models?library=ctransformers&sort=downloads).
+
+## Installation
+
+```sh
+pip install ctransformers
+```
+
+## Using existing models
+
+Load model directly:
+
+```py
+from ctransformers import AutoModelForCausalLM, AutoTokenizer
+
+model = AutoModelForCausalLM.from_pretrained("marella/gpt-2-ggml", hf=True)
+tokenizer = AutoTokenizer.from_pretrained(model)
+```
+
+Use a pipeline as a high-level helper:
+
+```py
+from transformers import pipeline
+
+pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
+```
+
+[Run in Google Colab](https://colab.research.google.com/drive/1FVSLfTJ2iBbQ1oU2Rqz0MkpJbaB_5Got)
+
+If you want to see how to load a specific model, you can click `Use in CTransformers` and you will be given a working snippet that you can load it!
+
+## Additional resources
+
+- CTransformers [library](https://github.com/marella/ctransformers).

docs/hub/models-libraries.md

@@ -12,6 +12,7 @@ The table below summarizes the supported libraries and their level of integratio
 | [AllenNLP](https://github.com/allenai/allennlp)                             | An open-source NLP research library, built on PyTorch.                               | ✅ | ✅ | ✅ | ❌ |
 | [Asteroid](https://github.com/asteroid-team/asteroid)                       | Pytorch-based audio source separation toolkit                                        | ✅ | ✅ | ✅ | ❌ |
 | [BERTopic](https://github.com/MaartenGr/BERTopic)                           | BERTopic is a topic modeling library for text and images                             | ✅ | ✅ | ✅ | ✅ | 
+| [CTransformers](https://github.com/marella/ctransformers)                   | Python bindings for the Transformer models implemented in C/C++ using GGML library   | ❌ | ❌ | ✅ | ❌ |
 | [docTR](https://github.com/mindee/doctr)                                    | Models and datasets for OCR-related tasks in PyTorch & TensorFlow                    | ✅ | ✅ | ✅ | ❌ |
 | [ESPnet](https://github.com/espnet/espnet)                                  | End-to-end speech processing toolkit (e.g. TTS)                                      | ✅ | ✅ | ✅ | ❌ |
 | [fastai](https://github.com/fastai/fastai)                                  | Library to train fast and accurate models with state-of-the-art outputs.             | ✅ | ✅ | ✅ | ✅ |

js/src/lib/interfaces/Libraries.ts

@@ -12,6 +12,7 @@ export enum ModelLibrary {
 	"allennlp"              = "allenNLP",
 	"asteroid"              = "Asteroid",
 	"bertopic"              = "BERTopic",
+	"ctransformers"         = "CTransformers",
 	"diffusers"             = "Diffusers",
 	"doctr"                 = "docTR",
 	"espnet"                = "ESPnet",
@@ -122,6 +123,19 @@ const bertopic = (model: ModelData) =>
 
 model = BERTopic.load("${model.id}")`];
 
+const ctransformers = (model: ModelData) =>
+	[
+		`# Load model directly
+from ctransformers import AutoModelForCausalLM, AutoTokenizer
+
+model = AutoModelForCausalLM.from_pretrained("${model.id}", hf=True)
+tokenizer = AutoTokenizer.from_pretrained(model)`,
+		`# Use a pipeline as a high-level helper
+from transformers import pipeline
+
+pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)`,
+	];
+
 const diffusers_default = (model: ModelData) =>
 	[`from diffusers import DiffusionPipeline
 
@@ -577,6 +591,12 @@ export const MODEL_LIBRARIES_UI_ELEMENTS: Partial<Record<ModelLibraryKey, Librar
 		repoUrl:  "https://github.com/MaartenGr/BERTopic",
 		snippets: bertopic,
 	},
+	"ctransformers": {
+		btnLabel: "CTransformers",
+		repoName: "ctransformers",
+		repoUrl:  "https://github.com/marella/ctransformers",
+		snippets: ctransformers,
+	},
 	"diffusers": {
 		btnLabel: "Diffusers",
 		repoName: "🤗/diffusers",
@@ -734,4 +754,3 @@ export const MODEL_LIBRARIES_UI_ELEMENTS: Partial<Record<ModelLibraryKey, Librar
 		snippets: pythae,
 	},
 } as const;
-

tasks/src/const.ts

@@ -33,7 +33,7 @@ export const TASKS_MODEL_LIBRARIES: Record<PipelineType, ModelLibraryKey[]> = {
 	"tabular-regression":             ["sklearn"],
 	"tabular-to-text":                ["transformers"],
 	"text-classification":            ["adapter-transformers", "spacy", "transformers", "transformers.js"],
-	"text-generation":                ["transformers", "transformers.js"],
+	"text-generation":                ["ctransformers", "transformers", "transformers.js"],
 	"text-retrieval":                 [],
 	"text-to-image":                  [],
 	"text-to-speech":                 ["espnet", "tensorflowtts"],