torchquery 2.1.1__tar.gz → 2.2.0__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.
Files changed (11) hide show
  1. {torchquery-2.1.1/torchquery.egg-info → torchquery-2.2.0}/PKG-INFO +18 -4
  2. {torchquery-2.1.1 → torchquery-2.2.0}/README.md +17 -3
  3. {torchquery-2.1.1 → torchquery-2.2.0}/pyproject.toml +1 -1
  4. {torchquery-2.1.1 → torchquery-2.2.0}/setup.py +1 -1
  5. {torchquery-2.1.1 → torchquery-2.2.0/torchquery.egg-info}/PKG-INFO +18 -4
  6. {torchquery-2.1.1 → torchquery-2.2.0}/LICENSE +0 -0
  7. {torchquery-2.1.1 → torchquery-2.2.0}/setup.cfg +0 -0
  8. {torchquery-2.1.1 → torchquery-2.2.0}/torchquery.egg-info/SOURCES.txt +0 -0
  9. {torchquery-2.1.1 → torchquery-2.2.0}/torchquery.egg-info/dependency_links.txt +0 -0
  10. {torchquery-2.1.1 → torchquery-2.2.0}/torchquery.egg-info/requires.txt +0 -0
  11. {torchquery-2.1.1 → torchquery-2.2.0}/torchquery.egg-info/top_level.txt +0 -0
{torchquery-2.1.1/torchquery.egg-info → torchquery-2.2.0}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: torchquery
3
- Version: 2.1.1
3
+ Version: 2.2.0
4
4
  Summary: High-performance SDC detection and neural healing for billion-scale tensors.
5
5
  Home-page: https://github.com/powerofaisinstudy-debug/torchquery
6
6
  Author: Sundaram Gupta
@@ -43,7 +43,21 @@ Dynamic: requires-python
43
43
 
44
44
  ---
45
45
 
46
- ## 📦 Installation
46
+ <p align="center">
47
+ <img src="https://raw.githubusercontent.com/powerofaisinstudy-debug/torchquery/main/chl.png" width="600">
48
+ </p>
49
+
50
+ ### Visualizing Silent Data Corruption (SDC)
51
+
52
+ Hardware glitches—like cosmic rays or VRAM overclocks—can cause random bit-flips. These create massive statistical outliers or `NaNs` in your tensor data.
53
+
54
+ [Image Link to Image_5.png]
55
+
56
+ **TorchQuery** acts as a `Neural Shield` that sweeps your multidimensional arrays. It identifies values that can lead to exploding gradients (`3e38`) or numerical instability (`NaN`), "healing" them before they propagate.
57
+
58
+ **Pre-Sweep State:**
59
+ * `NaN` (Not a Number): Corrupts entire model during backpropagation.
60
+ * `3e38`: Causes exploding gradients, destroying training stability.
47
61
 
48
- ```bash
49
- pip install torchquery
62
+ **Post-Sweep State:**
63
+ * Invalid data is removed, leaving behind validated tensor values.
{torchquery-2.1.1 → torchquery-2.2.0}/README.md RENAMED
@@ -20,7 +20,21 @@
20
20
 
21
21
  ---
22
22
 
23
- ## 📦 Installation
23
+ <p align="center">
24
+ <img src="https://raw.githubusercontent.com/powerofaisinstudy-debug/torchquery/main/chl.png" width="600">
25
+ </p>
26
+
27
+ ### Visualizing Silent Data Corruption (SDC)
28
+
29
+ Hardware glitches—like cosmic rays or VRAM overclocks—can cause random bit-flips. These create massive statistical outliers or `NaNs` in your tensor data.
30
+
31
+ [Image Link to Image_5.png]
32
+
33
+ **TorchQuery** acts as a `Neural Shield` that sweeps your multidimensional arrays. It identifies values that can lead to exploding gradients (`3e38`) or numerical instability (`NaN`), "healing" them before they propagate.
34
+
35
+ **Pre-Sweep State:**
36
+ * `NaN` (Not a Number): Corrupts entire model during backpropagation.
37
+ * `3e38`: Causes exploding gradients, destroying training stability.
24
38
 
25
- ```bash
26
- pip install torchquery
39
+ **Post-Sweep State:**
40
+ * Invalid data is removed, leaving behind validated tensor values.
{torchquery-2.1.1 → torchquery-2.2.0}/pyproject.toml RENAMED
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
4
4
 
5
5
  [project]
6
6
  name = "torchquery"
7
- version = "2.1.1"
7
+ version = "2.2.0"
8
8
  authors = [
9
9
  { name="Sundaram Gupta"},
10
10
  ]
{torchquery-2.1.1 → torchquery-2.2.0}/setup.py RENAMED
@@ -8,7 +8,7 @@ with open("README.md", "r", encoding="utf-8") as fh:
8
8
 
9
9
  setup(
10
10
  name="torchquery",
11
- version="2.1.1",
11
+ version="2.2.0",
12
12
  author="Sundaram Gupta",
13
13
  author_email="your-email@example.com", # Update this with your email
14
14
  description="A high-performance SDC detection and neural healing engine for billion-scale tensors.",
{torchquery-2.1.1 → torchquery-2.2.0/torchquery.egg-info}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: torchquery
3
- Version: 2.1.1
3
+ Version: 2.2.0
4
4
  Summary: High-performance SDC detection and neural healing for billion-scale tensors.
5
5
  Home-page: https://github.com/powerofaisinstudy-debug/torchquery
6
6
  Author: Sundaram Gupta
@@ -43,7 +43,21 @@ Dynamic: requires-python
43
43
 
44
44
  ---
45
45
 
46
- ## 📦 Installation
46
+ <p align="center">
47
+ <img src="https://raw.githubusercontent.com/powerofaisinstudy-debug/torchquery/main/chl.png" width="600">
48
+ </p>
49
+
50
+ ### Visualizing Silent Data Corruption (SDC)
51
+
52
+ Hardware glitches—like cosmic rays or VRAM overclocks—can cause random bit-flips. These create massive statistical outliers or `NaNs` in your tensor data.
53
+
54
+ [Image Link to Image_5.png]
55
+
56
+ **TorchQuery** acts as a `Neural Shield` that sweeps your multidimensional arrays. It identifies values that can lead to exploding gradients (`3e38`) or numerical instability (`NaN`), "healing" them before they propagate.
57
+
58
+ **Pre-Sweep State:**
59
+ * `NaN` (Not a Number): Corrupts entire model during backpropagation.
60
+ * `3e38`: Causes exploding gradients, destroying training stability.
47
61
 
48
- ```bash
49
- pip install torchquery
62
+ **Post-Sweep State:**
63
+ * Invalid data is removed, leaving behind validated tensor values.
{torchquery-2.1.1 → torchquery-2.2.0}/LICENSE RENAMED
File without changes
{torchquery-2.1.1 → torchquery-2.2.0}/setup.cfg RENAMED
File without changes