shanto268 commited on
Commit
25bdbb1
·
verified ·
1 Parent(s): 6dd46d3

README: the models are QCap, QCap-Lite and QCap-Fast (folders qcap/, qcap-lite/, qcap-fast/; API ids qcap, qcap-fast)

Browse files
Files changed (1) hide show
  1. README.md +27 -27
README.md CHANGED
@@ -10,7 +10,7 @@ datasets:
10
  <img src="https://github.com/LFL-Lab/SQuADDS/blob/master/docs/_static/images/squadds_logo_dark_name.png?raw=true" width="80%" alt="SQuADDS logo" />
11
  </center>
12
 
13
- # PlanarCapacitance-flagship: our model trained on all of our Q3D data
14
 
15
  The model of our paper *A physics-corrected graph neural network for the capacitance matrix of any planar superconducting circuit layout*, trained on all 45,771 designs of our Q3D dataset, the test designs included. It reads a GDS layout and returns the full Maxwell capacitance matrix of every conductor in it. Because it has seen the test designs, it has no held-out score of its own. Its held-out accuracy, from the same model trained without the test designs, is in the paper.
16
 
@@ -18,7 +18,7 @@ The model of our paper *A physics-corrected graph neural network for the capacit
18
  - **Paper:** [lfl-lab.github.io/SQuADDS/ml-paper/paper](https://lfl-lab.github.io/SQuADDS/ml-paper/paper/)
19
  - **Code:** [github.com/LFL-Lab/SQuADDS](https://github.com/LFL-Lab/SQuADDS)
20
  - **Data:** [Q3D dataset](https://huggingface.co/datasets/SQuADDS/SQuADDS_PlanarCapacitance) and [synthetic dataset](https://huggingface.co/datasets/SQuADDS/SQuADDS_PlanarCapacitance-Synthetic), part of the [SQuADDS datasets](https://huggingface.co/collections/SQuADDS/squadds-datasets-6a65152b857f8efa5620880a)
21
- - **Models:** [flagship models](https://huggingface.co/SQuADDS/PlanarCapacitance-flagship), part of the [SQuADDS models](https://huggingface.co/collections/SQuADDS/squadds-models-6af8b7da67a22df62af374c6)
22
  - **API:** [SQuADDS ML inference API](https://huggingface.co/spaces/SQuADDS/squadds-ml-inference-api)
23
  - **Point of contact:** [Sadman Ahmed Shanto](mailto:shanto@usc.edu)
24
 
@@ -27,28 +27,28 @@ The model of our paper *A physics-corrected graph neural network for the capacit
27
  ```python
28
  from squadds.ml.capacitance import CapacitancePredictor, load_model
29
 
30
- res = CapacitancePredictor(load_model("flagship")).predict("chip.gds") # Accurate
31
- fast = CapacitancePredictor(load_model("fast")).predict("chip.gds") # Fast
32
  res.names, res.matrix_fF, res.to_sim_results()
33
  ```
34
 
35
  ## Hosted API
36
 
37
- The [SQuADDS ML inference API](https://huggingface.co/spaces/SQuADDS/squadds-ml-inference-api) serves these models over HTTP as `planar_capacitance_flagship` (Accurate) and `planar_capacitance_fast`. They go live there when the paper is published. Until then, run them locally with SQuADDS as above.
38
 
39
- - `GET /v1/models/planar_capacitance_flagship` describes the model, with its inputs, limits and an example.
40
- - `GET /v1/models/planar_capacitance_flagship/schema?format=tool` gives a ready tool definition for AI agents.
41
- - `POST /v1/models/planar_capacitance_flagship/predict` takes the GDS file as the multipart field `file`, or base64 encoded in `inputs.file_base64` of a JSON body. The optional inputs are `ground_layer` and `conductor_layer` (`1/0` and `1/10` by default), `eps_r` and `thickness_um` (11.45 and 350 µm), and `ground_pieces` when ground and conductors share a layer.
42
 
43
  ```bash
44
- curl -X POST https://squadds-squadds-ml-inference-api.hf.space/v1/models/planar_capacitance_flagship/predict \
45
  -F file=@chip.gds -F 'inputs={"ground_layer": "1/0", "conductor_layer": "1/10"}'
46
  ```
47
 
48
  ```python
49
  from squadds.ml import client
50
 
51
- r = client.predict("planar_capacitance_flagship", file="chip.gds")
52
  r["outputs"]["names"], r["outputs"]["matrix_fF"]
53
  ```
54
 
@@ -56,7 +56,7 @@ The response for the interdigital example layout:
56
 
57
  ```json
58
  {
59
- "model_id": "planar_capacitance_flagship",
60
  "outputs": {
61
  "names": ["G", "C1", "C2"],
62
  "matrix_fF": [[31.727, -4.264, -5.711], [-4.264, 6.598, -2.285], [-5.711, -2.285, 8.045]],
@@ -70,22 +70,22 @@ The API takes up to 70 conductors plus ground and 20 MB per file. It runs one pr
70
 
71
  ## Models
72
 
73
- This repository holds all three models in `accurate/`, `balanced/` and `fast/`. SQuADDS loads Accurate and Fast.
74
 
75
  | model | how it works | error on 39 new layouts, % (MAE / RMSE, fF) | median time |
76
  |---|---|---|---|
77
- | Accurate (default) | the extrapolated prior from two coarse method-of-moments solves, corrected by the networks | 1.05 (0.132 / 0.188) | 1.1 s |
78
- | Balanced | the same recipe on two coarser solves | 2.044 (0.258 / 0.348) | 0.56 s |
79
- | Fast | a learned prior replaces the solves, then the networks correct it | 3.175 (0.583 / 0.824) | 0.24 s |
80
 
81
  The 39 layouts are generated rings with a neighbour (3 or 4 conductors) that no part of any model saw in training. On them the extrapolated prior alone scores 1.765 % (MAE 0.275 / RMSE 0.376 fF). Times are on two CPUs of an Intel Xeon E5-2640 v4 cluster node.
82
 
83
- Fast is for quick screening of the component families it was trained on. Use Accurate otherwise.
84
 
85
  ## Conventions
86
 
87
  - Capacitance is in fF. Ground is row and column 0, and nets are never merged.
88
- - Trained on a 350 µm silicon substrate (eps_r 11.45) with zero-thickness metal. Accurate takes other substrates with `predict(..., eps_r=..., thickness_um=...)`, as in Sec. IV C of the paper. Fast only rescales by permittivity there and warns.
89
  - Ground on GDS layer 1/0 and conductors on 1/10 by default. Every result carries its physical checks, and a failed check is raised as a warning.
90
 
91
  ## Citation
@@ -125,13 +125,13 @@ SQuADDS checks every file against the manifest in the package (`squadds/ml/capac
125
 
126
  | file | sha256 |
127
  |---|---|
128
- | `accurate/acc_v2m_sf_s0.pt` | `cc69c5818f5028c26177159b88239116520abbb290743d6a6c091657e8dbf1c5` |
129
- | `accurate/acc_v2m_sf_s1.pt` | `2f88606dbec5c599550d088213a4c37af3b3ad5134ebad2e60519fb534553a57` |
130
- | `accurate/acc_v2m_sf_s2.pt` | `d13bafc6b2eb7cadcd97a387d791c0210976db1a9d618f9acc1b614e3701daed` |
131
- | `balanced/bal_v2m_sf_s0.pt` | `aa742ab93b1e3ffcd86a753bda5cc4a68812acbe8140a6b0e311c11d1f36b4b2` |
132
- | `balanced/bal_v2m_sf_s1.pt` | `486466473eba61c090f4e3fa7988bf880fc183570c5d02bd78f4c1356691f926` |
133
- | `balanced/bal_v2m_sf_s2.pt` | `1b55ded04fd624585eeeccf2ab6e4a8e9354a9b96694131e1737dde43ad3ff58` |
134
- | `fast/acc_v1_s0.pt` | `9dff20ade26619e7713c2d145dada65e4554a623780290673cfe9452cf9b64ba` |
135
- | `fast/acc_v1_s1.pt` | `f1ad2ca065dd650ccabcad73a3c6ad2ac1042f4f68bb828893bbdecc47411ffe` |
136
- | `fast/acc_v1_s2.pt` | `8cbada211724df5cc05a7b4fdd232607de047b248a43993d3844bd68ff8060c4` |
137
- | `fast/sall_all_s0.pt` | `132d6f389eda798d31bd5654a96e24cce18fa750134a9fb2a9725eafed339dac` |
 
10
  <img src="https://github.com/LFL-Lab/SQuADDS/blob/master/docs/_static/images/squadds_logo_dark_name.png?raw=true" width="80%" alt="SQuADDS logo" />
11
  </center>
12
 
13
+ # QCap: our model trained on all of our Q3D data
14
 
15
  The model of our paper *A physics-corrected graph neural network for the capacitance matrix of any planar superconducting circuit layout*, trained on all 45,771 designs of our Q3D dataset, the test designs included. It reads a GDS layout and returns the full Maxwell capacitance matrix of every conductor in it. Because it has seen the test designs, it has no held-out score of its own. Its held-out accuracy, from the same model trained without the test designs, is in the paper.
16
 
 
18
  - **Paper:** [lfl-lab.github.io/SQuADDS/ml-paper/paper](https://lfl-lab.github.io/SQuADDS/ml-paper/paper/)
19
  - **Code:** [github.com/LFL-Lab/SQuADDS](https://github.com/LFL-Lab/SQuADDS)
20
  - **Data:** [Q3D dataset](https://huggingface.co/datasets/SQuADDS/SQuADDS_PlanarCapacitance) and [synthetic dataset](https://huggingface.co/datasets/SQuADDS/SQuADDS_PlanarCapacitance-Synthetic), part of the [SQuADDS datasets](https://huggingface.co/collections/SQuADDS/squadds-datasets-6a65152b857f8efa5620880a)
21
+ - **Models:** [QCap models](https://huggingface.co/SQuADDS/QCap) (QCap, QCap-Lite and QCap-Fast), part of the [SQuADDS models](https://huggingface.co/collections/SQuADDS/squadds-models-6af8b7da67a22df62af374c6)
22
  - **API:** [SQuADDS ML inference API](https://huggingface.co/spaces/SQuADDS/squadds-ml-inference-api)
23
  - **Point of contact:** [Sadman Ahmed Shanto](mailto:shanto@usc.edu)
24
 
 
27
  ```python
28
  from squadds.ml.capacitance import CapacitancePredictor, load_model
29
 
30
+ res = CapacitancePredictor(load_model("QCap")).predict("chip.gds") # QCap
31
+ fast = CapacitancePredictor(load_model("QCap-Fast")).predict("chip.gds") # QCap-Fast
32
  res.names, res.matrix_fF, res.to_sim_results()
33
  ```
34
 
35
  ## Hosted API
36
 
37
+ The [SQuADDS ML inference API](https://huggingface.co/spaces/SQuADDS/squadds-ml-inference-api) serves these models over HTTP as `qcap` (QCap) and `qcap-fast` (QCap-Fast). They go live there when the paper is published. Until then, run them locally with SQuADDS as above.
38
 
39
+ - `GET /v1/models/qcap` describes the model, with its inputs, limits and an example.
40
+ - `GET /v1/models/qcap/schema?format=tool` gives a ready tool definition for AI agents.
41
+ - `POST /v1/models/qcap/predict` takes the GDS file as the multipart field `file`, or base64 encoded in `inputs.file_base64` of a JSON body. The optional inputs are `ground_layer` and `conductor_layer` (`1/0` and `1/10` by default), `eps_r` and `thickness_um` (11.45 and 350 µm), and `ground_pieces` when ground and conductors share a layer.
42
 
43
  ```bash
44
+ curl -X POST https://squadds-squadds-ml-inference-api.hf.space/v1/models/qcap/predict \
45
  -F file=@chip.gds -F 'inputs={"ground_layer": "1/0", "conductor_layer": "1/10"}'
46
  ```
47
 
48
  ```python
49
  from squadds.ml import client
50
 
51
+ r = client.predict("qcap", file="chip.gds")
52
  r["outputs"]["names"], r["outputs"]["matrix_fF"]
53
  ```
54
 
 
56
 
57
  ```json
58
  {
59
+ "model_id": "qcap",
60
  "outputs": {
61
  "names": ["G", "C1", "C2"],
62
  "matrix_fF": [[31.727, -4.264, -5.711], [-4.264, 6.598, -2.285], [-5.711, -2.285, 8.045]],
 
70
 
71
  ## Models
72
 
73
+ This repository holds all three models in `qcap/`, `qcap-lite/` and `qcap-fast/`. SQuADDS loads QCap and QCap-Fast.
74
 
75
  | model | how it works | error on 39 new layouts, % (MAE / RMSE, fF) | median time |
76
  |---|---|---|---|
77
+ | QCap (default) | the extrapolated prior from two coarse method-of-moments solves, corrected by the networks | 1.05 (0.132 / 0.188) | 1.1 s |
78
+ | QCap-Lite | the same recipe on two coarser solves | 2.044 (0.258 / 0.348) | 0.56 s |
79
+ | QCap-Fast | a learned prior replaces the solves, then the networks correct it | 3.175 (0.583 / 0.824) | 0.24 s |
80
 
81
  The 39 layouts are generated rings with a neighbour (3 or 4 conductors) that no part of any model saw in training. On them the extrapolated prior alone scores 1.765 % (MAE 0.275 / RMSE 0.376 fF). Times are on two CPUs of an Intel Xeon E5-2640 v4 cluster node.
82
 
83
+ QCap-Fast is for quick screening of the component families it was trained on. Use QCap otherwise.
84
 
85
  ## Conventions
86
 
87
  - Capacitance is in fF. Ground is row and column 0, and nets are never merged.
88
+ - Trained on a 350 µm silicon substrate (eps_r 11.45) with zero-thickness metal. QCap takes other substrates with `predict(..., eps_r=..., thickness_um=...)`, as in Sec. IV C of the paper. QCap-Fast only rescales by permittivity there and warns.
89
  - Ground on GDS layer 1/0 and conductors on 1/10 by default. Every result carries its physical checks, and a failed check is raised as a warning.
90
 
91
  ## Citation
 
125
 
126
  | file | sha256 |
127
  |---|---|
128
+ | `qcap/acc_v2m_sf_s0.pt` | `cc69c5818f5028c26177159b88239116520abbb290743d6a6c091657e8dbf1c5` |
129
+ | `qcap/acc_v2m_sf_s1.pt` | `2f88606dbec5c599550d088213a4c37af3b3ad5134ebad2e60519fb534553a57` |
130
+ | `qcap/acc_v2m_sf_s2.pt` | `d13bafc6b2eb7cadcd97a387d791c0210976db1a9d618f9acc1b614e3701daed` |
131
+ | `qcap-lite/bal_v2m_sf_s0.pt` | `aa742ab93b1e3ffcd86a753bda5cc4a68812acbe8140a6b0e311c11d1f36b4b2` |
132
+ | `qcap-lite/bal_v2m_sf_s1.pt` | `486466473eba61c090f4e3fa7988bf880fc183570c5d02bd78f4c1356691f926` |
133
+ | `qcap-lite/bal_v2m_sf_s2.pt` | `1b55ded04fd624585eeeccf2ab6e4a8e9354a9b96694131e1737dde43ad3ff58` |
134
+ | `qcap-fast/acc_v1_s0.pt` | `9dff20ade26619e7713c2d145dada65e4554a623780290673cfe9452cf9b64ba` |
135
+ | `qcap-fast/acc_v1_s1.pt` | `f1ad2ca065dd650ccabcad73a3c6ad2ac1042f4f68bb828893bbdecc47411ffe` |
136
+ | `qcap-fast/acc_v1_s2.pt` | `8cbada211724df5cc05a7b4fdd232607de047b248a43993d3844bd68ff8060c4` |
137
+ | `qcap-fast/sall_all_s0.pt` | `132d6f389eda798d31bd5654a96e24cce18fa750134a9fb2a9725eafed339dac` |