softcane commited on
Commit
8260d85
·
verified ·
1 Parent(s): 4a1479c

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -11
README.md CHANGED
@@ -13,17 +13,15 @@ tags:
13
 
14
  # SpotVortex TFT Dual-Head Risk Model (ONNX)
15
 
16
- Production ONNX artifact used by SpotVortex for spot instance risk inference. Outputs dual risk scores capacity and runtime per instance type and availability zone.
17
 
18
  ---
19
-
20
  ## Data Source
21
 
22
  Training data is sourced from **SpotLake** spot market telemetry:
23
  🔗 [https://spotlake.ddps.cloud/](https://spotlake.ddps.cloud/)
24
 
25
  ---
26
-
27
  ## Cloud Scope
28
 
29
  | Dimension | Value |
@@ -32,7 +30,6 @@ Training data is sourced from **SpotLake** spot market telemetry:
32
  | Instance Families | 57 (see below) |
33
 
34
  ### Supported Instance Families
35
-
36
  | Gen | Compute (`c`) | General Purpose (`m`) | Memory (`r`) | Burstable (`t`) |
37
  |:---:|---|---|---|---|
38
  | **5** | `c5` `c5a` `c5ad` `c5d` `c5n` | `m5` `m5a` `m5ad` `m5d` `m5dn` `m5n` `m5zn` | `r5` `r5a` `r5ad` `r5b` `r5d` `r5dn` `r5n` | `t2` `t3` `t3a` |
@@ -40,9 +37,7 @@ Training data is sourced from **SpotLake** spot market telemetry:
40
  | **7** | `c7a` `c7g` `c7gd` `c7gn` `c7i` `c7i-flex` | `m7a` `m7g` `m7gd` `m7i` `m7i-flex` | `r7a` `r7g` `r7gd` `r7i` `r7iz` | `t4g` |
41
 
42
  ---
43
-
44
  ## Artifact Contents
45
-
46
  | File | Description |
47
  |------|-------------|
48
  | `tft.onnx` | ONNX computation graph |
@@ -52,11 +47,9 @@ Training data is sourced from **SpotLake** spot market telemetry:
52
  | `LICENSE` | Non-commercial proprietary terms |
53
 
54
  ---
55
-
56
  ## Model Interface
57
 
58
  ### Input
59
-
60
  | Property | Value |
61
  |----------|-------|
62
  | Tensor type | `float32` |
@@ -64,16 +57,13 @@ Training data is sourced from **SpotLake** spot market telemetry:
64
  | Features per timestep | 10 |
65
 
66
  ### Outputs
67
-
68
  | Tensor Name | Description |
69
  |-------------|-------------|
70
  | `capacity_score` | Predicted spot capacity availability risk |
71
  | `runtime_score` | Predicted spot instance interruption/runtime risk |
72
 
73
  ---
74
-
75
  ## Usage & Restrictions
76
-
77
  | Topic | Detail |
78
  |-------|--------|
79
  | Runtime integration | Use the SpotVortex agent repository |
 
13
 
14
  # SpotVortex TFT Dual-Head Risk Model (ONNX)
15
 
16
+ Production ONNX artifact used by SpotVortex for spot instance risk inference. Outputs dual risk scores; capacity and runtime per instance type and availability zone.
17
 
18
  ---
 
19
  ## Data Source
20
 
21
  Training data is sourced from **SpotLake** spot market telemetry:
22
  🔗 [https://spotlake.ddps.cloud/](https://spotlake.ddps.cloud/)
23
 
24
  ---
 
25
  ## Cloud Scope
26
 
27
  | Dimension | Value |
 
30
  | Instance Families | 57 (see below) |
31
 
32
  ### Supported Instance Families
 
33
  | Gen | Compute (`c`) | General Purpose (`m`) | Memory (`r`) | Burstable (`t`) |
34
  |:---:|---|---|---|---|
35
  | **5** | `c5` `c5a` `c5ad` `c5d` `c5n` | `m5` `m5a` `m5ad` `m5d` `m5dn` `m5n` `m5zn` | `r5` `r5a` `r5ad` `r5b` `r5d` `r5dn` `r5n` | `t2` `t3` `t3a` |
 
37
  | **7** | `c7a` `c7g` `c7gd` `c7gn` `c7i` `c7i-flex` | `m7a` `m7g` `m7gd` `m7i` `m7i-flex` | `r7a` `r7g` `r7gd` `r7i` `r7iz` | `t4g` |
38
 
39
  ---
 
40
  ## Artifact Contents
 
41
  | File | Description |
42
  |------|-------------|
43
  | `tft.onnx` | ONNX computation graph |
 
47
  | `LICENSE` | Non-commercial proprietary terms |
48
 
49
  ---
 
50
  ## Model Interface
51
 
52
  ### Input
 
53
  | Property | Value |
54
  |----------|-------|
55
  | Tensor type | `float32` |
 
57
  | Features per timestep | 10 |
58
 
59
  ### Outputs
 
60
  | Tensor Name | Description |
61
  |-------------|-------------|
62
  | `capacity_score` | Predicted spot capacity availability risk |
63
  | `runtime_score` | Predicted spot instance interruption/runtime risk |
64
 
65
  ---
 
66
  ## Usage & Restrictions
 
67
  | Topic | Detail |
68
  |-------|--------|
69
  | Runtime integration | Use the SpotVortex agent repository |