controlmt 0.1.2__tar.gz → 0.1.3__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -1,12 +1,15 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: controlmt
3
- Version: 0.1.2
3
+ Version: 0.1.3
4
4
  Summary: Python SDK for the ControlMT v2.3 Kannada↔English translator
5
5
  Author: Anand Kaman
6
6
  License: Apache-2.0
7
7
  Project-URL: Homepage, https://huggingface.co/anandkaman/controlmt-v2.3
8
+ Project-URL: Documentation, https://huggingface.co/anandkaman/controlmt-v2.3/blob/main/DEPLOYMENT.md
8
9
  Project-URL: Source, https://github.com/anandkaman/ControlMT
9
10
  Project-URL: Demo, https://huggingface.co/spaces/anandkaman/controlmt-demo
11
+ Project-URL: Quantized (int8), https://huggingface.co/anandkaman/controlmt-v2.3-int8
12
+ Project-URL: Issue Tracker, https://github.com/anandkaman/ControlMT/issues
10
13
  Keywords: translation,kannada,machine-translation,nlp,indic,controlmt
11
14
  Classifier: License :: OSI Approved :: Apache Software License
12
15
  Classifier: Operating System :: OS Independent
@@ -31,10 +34,19 @@ Requires-Dist: pytest>=7; extra == "test"
31
34
  Python SDK for **[ControlMT v2.3](https://huggingface.co/anandkaman/controlmt-v2.3)** —
32
35
  a compact 139M-parameter Kannada ↔ English translator.
33
36
 
37
+ ## Install
38
+
34
39
  ```bash
40
+ # CPU-only (smaller — ~200 MB torch wheel)
41
+ pip install torch --index-url https://download.pytorch.org/whl/cpu
42
+ pip install controlmt
43
+
44
+ # GPU (CUDA) — default; pulls ~2 GB CUDA torch wheel
35
45
  pip install controlmt
36
46
  ```
37
47
 
48
+ > The default `pip install controlmt` works on both CPU and GPU machines — but pulls the GPU-enabled torch wheel (~2 GB). If you don't have a CUDA GPU, install the CPU-only torch wheel first (one extra command, ~200 MB) and the model still runs at full speed via bf16 / int8 dynamic quantization.
49
+
38
50
  ## Quick start
39
51
 
40
52
  ```python
@@ -3,10 +3,19 @@
3
3
  Python SDK for **[ControlMT v2.3](https://huggingface.co/anandkaman/controlmt-v2.3)** —
4
4
  a compact 139M-parameter Kannada ↔ English translator.
5
5
 
6
+ ## Install
7
+
6
8
  ```bash
9
+ # CPU-only (smaller — ~200 MB torch wheel)
10
+ pip install torch --index-url https://download.pytorch.org/whl/cpu
11
+ pip install controlmt
12
+
13
+ # GPU (CUDA) — default; pulls ~2 GB CUDA torch wheel
7
14
  pip install controlmt
8
15
  ```
9
16
 
17
+ > The default `pip install controlmt` works on both CPU and GPU machines — but pulls the GPU-enabled torch wheel (~2 GB). If you don't have a CUDA GPU, install the CPU-only torch wheel first (one extra command, ~200 MB) and the model still runs at full speed via bf16 / int8 dynamic quantization.
18
+
10
19
  ## Quick start
11
20
 
12
21
  ```python
@@ -33,9 +33,12 @@ dependencies = [
33
33
  test = ["pytest>=7"]
34
34
 
35
35
  [project.urls]
36
- Homepage = "https://huggingface.co/anandkaman/controlmt-v2.3"
37
- Source = "https://github.com/anandkaman/ControlMT"
38
- Demo = "https://huggingface.co/spaces/anandkaman/controlmt-demo"
36
+ Homepage = "https://huggingface.co/anandkaman/controlmt-v2.3"
37
+ Documentation = "https://huggingface.co/anandkaman/controlmt-v2.3/blob/main/DEPLOYMENT.md"
38
+ Source = "https://github.com/anandkaman/ControlMT"
39
+ Demo = "https://huggingface.co/spaces/anandkaman/controlmt-demo"
40
+ "Quantized (int8)" = "https://huggingface.co/anandkaman/controlmt-v2.3-int8"
41
+ "Issue Tracker" = "https://github.com/anandkaman/ControlMT/issues"
39
42
 
40
43
  [tool.setuptools.dynamic]
41
44
  version = { attr = "controlmt._version.__version__" }
@@ -0,0 +1 @@
1
+ __version__ = "0.1.3"
@@ -1,12 +1,15 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: controlmt
3
- Version: 0.1.2
3
+ Version: 0.1.3
4
4
  Summary: Python SDK for the ControlMT v2.3 Kannada↔English translator
5
5
  Author: Anand Kaman
6
6
  License: Apache-2.0
7
7
  Project-URL: Homepage, https://huggingface.co/anandkaman/controlmt-v2.3
8
+ Project-URL: Documentation, https://huggingface.co/anandkaman/controlmt-v2.3/blob/main/DEPLOYMENT.md
8
9
  Project-URL: Source, https://github.com/anandkaman/ControlMT
9
10
  Project-URL: Demo, https://huggingface.co/spaces/anandkaman/controlmt-demo
11
+ Project-URL: Quantized (int8), https://huggingface.co/anandkaman/controlmt-v2.3-int8
12
+ Project-URL: Issue Tracker, https://github.com/anandkaman/ControlMT/issues
10
13
  Keywords: translation,kannada,machine-translation,nlp,indic,controlmt
11
14
  Classifier: License :: OSI Approved :: Apache Software License
12
15
  Classifier: Operating System :: OS Independent
@@ -31,10 +34,19 @@ Requires-Dist: pytest>=7; extra == "test"
31
34
  Python SDK for **[ControlMT v2.3](https://huggingface.co/anandkaman/controlmt-v2.3)** —
32
35
  a compact 139M-parameter Kannada ↔ English translator.
33
36
 
37
+ ## Install
38
+
34
39
  ```bash
40
+ # CPU-only (smaller — ~200 MB torch wheel)
41
+ pip install torch --index-url https://download.pytorch.org/whl/cpu
42
+ pip install controlmt
43
+
44
+ # GPU (CUDA) — default; pulls ~2 GB CUDA torch wheel
35
45
  pip install controlmt
36
46
  ```
37
47
 
48
+ > The default `pip install controlmt` works on both CPU and GPU machines — but pulls the GPU-enabled torch wheel (~2 GB). If you don't have a CUDA GPU, install the CPU-only torch wheel first (one extra command, ~200 MB) and the model still runs at full speed via bf16 / int8 dynamic quantization.
49
+
38
50
  ## Quick start
39
51
 
40
52
  ```python
@@ -1 +0,0 @@
1
- __version__ = "0.1.2"
File without changes