Add OmniScient results (fine-tuned Chronos-2 derivative, 360 Labs) (#36)

- Add OmniScient results (fine-tuned Chronos-2 derivative, 360 Labs) (ba2ae424ee35c98b33e50fd84fe0e71660271dbe)


Co-authored-by: Suryansh Srivastava <quickbrownfawkes@users.noreply.huggingface.co>

results/OmniScient/Australia_Solar/H/long/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "Australia_Solar/H/long",
+  "num_series": 1,
+  "num_windows": 15,
+  "num_variates": 3,
+  "prediction_length": 168,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "H",
+  "seasonality": 24,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 100.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/Australia_Solar/H/long/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8dfa84ebd5c66e6854bc032af4ede428bf31483e7cf7d491ab7ca031b677c16
+size 4536

results/OmniScient/Australia_Solar/H/long/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa78fb9680ee454a616dc3cdbb0aa6300fd9c1577838547c04179a22e26a4688
+size 123974

results/OmniScient/Australia_Solar/H/medium/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "Australia_Solar/H/medium",
+  "num_series": 1,
+  "num_windows": 35,
+  "num_variates": 3,
+  "prediction_length": 72,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "H",
+  "seasonality": 24,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 100.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/Australia_Solar/H/medium/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:864cf362e05515180c1442c4e29a986c21195bb4cfe2b112aae3dcde74c1e8f7
+size 8352

results/OmniScient/Australia_Solar/H/medium/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:95e511e62ddc58d3c284746451800c321b9593bdaf6aa52922f207d427b95e08
+size 124157

results/OmniScient/Australia_Solar/H/short/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "Australia_Solar/H/short",
+  "num_series": 1,
+  "num_windows": 105,
+  "num_variates": 3,
+  "prediction_length": 24,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "H",
+  "seasonality": 24,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 100.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/Australia_Solar/H/short/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d6f516602931ae2bc5907352232e8158a3e2577e8717b0ba3f4f8793b557ba4
+size 20777

results/OmniScient/Australia_Solar/H/short/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7fbaac9378f849871237be57afbce132343e20ea1e01ca4b604a93072df65e9
+size 124139

results/OmniScient/Auto_Production_SF/M/short/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "Auto_Production_SF/M/short",
+  "num_series": 1,
+  "num_windows": 5,
+  "num_variates": 1,
+  "prediction_length": 12,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "M",
+  "seasonality": 12,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/Auto_Production_SF/M/short/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac6033af80f15d07e0153494845db5be49b4a669bec0cec43cbb985f5c3865df
+size 1823

results/OmniScient/Auto_Production_SF/M/short/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86335c92fc30cd7799cca1f1a492529e2c1722387262bf74f0052a16e9183723
+size 1325

results/OmniScient/CPHL/15T/long/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "CPHL/15T/long",
+  "num_series": 2,
+  "num_windows": 5,
+  "num_variates": 1,
+  "prediction_length": 288,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "15T",
+  "seasonality": 96,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/CPHL/15T/long/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c82ae24e3ea50382bcfcee833fbf469487e5c513d6ed3e4f51876f1c2b8560a7
+size 2174

results/OmniScient/CPHL/15T/long/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b0f60850708c816954793818fb406dcae572aa21e11841d7450b7b0240e77fe
+size 37969

results/OmniScient/CPHL/15T/medium/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "CPHL/15T/medium",
+  "num_series": 2,
+  "num_windows": 15,
+  "num_variates": 1,
+  "prediction_length": 96,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "15T",
+  "seasonality": 96,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/CPHL/15T/medium/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89b9e5a4772c53224d7aeeec2a13d4816f573deb347f28f9ddd7f460dd4a56c4
+size 3470

results/OmniScient/CPHL/15T/medium/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:315af65325cc4cb15e95c8861a0976c7774f21ade36117dc4314b3611d4f73e8
+size 38731

results/OmniScient/CPHL/15T/short/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "CPHL/15T/short",
+  "num_series": 2,
+  "num_windows": 120,
+  "num_variates": 1,
+  "prediction_length": 12,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "15T",
+  "seasonality": 96,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/CPHL/15T/short/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:790a490e40bba3f99ac4c72c75264595de6b72c24799ed414245aef7d566df3d
+size 15738

results/OmniScient/CPHL/15T/short/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e764a947e6d6b1c30b19e18fd12b7a23a11a726203bd65a6a7df0c5d2190d9fe
+size 39813

results/OmniScient/CPHL/30T/long/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "CPHL/30T/long",
+  "num_series": 2,
+  "num_windows": 10,
+  "num_variates": 1,
+  "prediction_length": 144,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "30T",
+  "seasonality": 48,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/CPHL/30T/long/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e2d5c437092f59918b1f6081faf0cfca15b9ed6deca8b28d366888135a4c088
+size 2837

results/OmniScient/CPHL/30T/long/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93a23cacd64698caff2d6978ce256d61a9bd534268796860f08ddd981be28066
+size 40330

results/OmniScient/CPHL/30T/medium/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "CPHL/30T/medium",
+  "num_series": 2,
+  "num_windows": 30,
+  "num_variates": 1,
+  "prediction_length": 48,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "30T",
+  "seasonality": 48,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/CPHL/30T/medium/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f677b3d6ad41615be4f3c3d4bd5a6b77af5c708c0a7b08a0edabe27eb456eb4
+size 5464

results/OmniScient/CPHL/30T/medium/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09465f5302c0b2cd1cd123f52b60b4bf18d73752a476d8f8ef269f67504755e6
+size 40711

results/OmniScient/CPHL/30T/short/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "CPHL/30T/short",
+  "num_series": 2,
+  "num_windows": 120,
+  "num_variates": 1,
+  "prediction_length": 12,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "30T",
+  "seasonality": 48,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/CPHL/30T/short/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a881aa093e81f5968285fa9d9a4e1cf68f315ab0d829459ef789c99100aebb82
+size 15887

results/OmniScient/CPHL/30T/short/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34e66d4d120e35489dcccf665ea64c466f26e9ee1076f27819dc8b6e44c4abfd
+size 41338

results/OmniScient/CPHL/H/long/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "CPHL/H/long",
+  "num_series": 4,
+  "num_windows": 2,
+  "num_variates": 1,
+  "prediction_length": 336,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "H",
+  "seasonality": 24,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/CPHL/H/long/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c35832fcf9ed628c7818bd9336aa06446088611131bc52adbc1bf27fd3127451
+size 2034

results/OmniScient/CPHL/H/long/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94879a4f7d0afdbeaa5c7fb83327365da05ca45dc9a25dc2f761a8c900fc6dd4
+size 37065

results/OmniScient/CPHL/H/medium/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "CPHL/H/medium",
+  "num_series": 4,
+  "num_windows": 4,
+  "num_variates": 1,
+  "prediction_length": 168,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "H",
+  "seasonality": 24,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/CPHL/H/medium/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b51be7a73c3cb794253ac5e980f6e553a1656c3ac8f4ed8e56861b4feb199e0
+size 2561

results/OmniScient/CPHL/H/medium/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0f98002422c79f4ce71c75793b2fa354f14afef682ad97c0fc36597bf4062a3
+size 37824

results/OmniScient/CPHL/H/short/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "CPHL/H/short",
+  "num_series": 4,
+  "num_windows": 28,
+  "num_variates": 1,
+  "prediction_length": 24,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "H",
+  "seasonality": 24,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/CPHL/H/short/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4218907b75a02a3469a32995d442d759b9cac94d5118baa287f4f0f43f62431b
+size 8503

results/OmniScient/CPHL/H/short/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13af4144c4b7d880b088a7dabc6c35e24d0537f4ea2ecd0f096b1951c875bd02
+size 38982

results/OmniScient/Coastal_T_S/15T/long/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "Coastal_T_S/15T/long",
+  "num_series": 5,
+  "num_windows": 5,
+  "num_variates": 3,
+  "prediction_length": 288,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "15T",
+  "seasonality": 96,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/Coastal_T_S/15T/long/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbebc7bff8f69110dab01d8d8785517b04276516f5654e179510fc2b630daef4
+size 6285

results/OmniScient/Coastal_T_S/15T/long/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0aa1020d020466b522da6e6b91aed836f0215e9f2c3c7c41072549a946f7e79d
+size 112530

results/OmniScient/Coastal_T_S/15T/medium/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "Coastal_T_S/15T/medium",
+  "num_series": 5,
+  "num_windows": 15,
+  "num_variates": 3,
+  "prediction_length": 96,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "15T",
+  "seasonality": 96,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/Coastal_T_S/15T/medium/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2001da48c94ea1e145954bbc4641280859be91798b77bc21319785228e018e1d
+size 14843

results/OmniScient/Coastal_T_S/15T/medium/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:374af09df7ef4be0a036d2c27233f43fc51c52c3b030a964ade23555c7070c99
+size 121652

results/OmniScient/Coastal_T_S/15T/short/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "Coastal_T_S/15T/short",
+  "num_series": 5,
+  "num_windows": 120,
+  "num_variates": 3,
+  "prediction_length": 12,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "15T",
+  "seasonality": 96,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/Coastal_T_S/15T/short/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dda08eba88f20d705af0e5e957835d63833c610cd00fcd3426f6f1812c5f7de6
+size 103062

results/OmniScient/Coastal_T_S/15T/short/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2653a2c8b1f84b402078415249102f4f632db0ca015b4b68d6014ffe1bee36b9
+size 144936

results/OmniScient/Coastal_T_S/20T/long/config.json

@@ -0,0 +1,37 @@
+{
+  "dataset_config": "Coastal_T_S/20T/long",
+  "num_series": 1,
+  "num_windows": 5,
+  "num_variates": 3,
+  "prediction_length": 216,
+  "num_quantiles": 9,
+  "quantile_levels": [
+    0.1,
+    0.2,
+    0.3,
+    0.4,
+    0.5,
+    0.6,
+    0.7,
+    0.8,
+    0.9
+  ],
+  "freq": "20T",
+  "seasonality": 72,
+  "context_length": 8192,
+  "metric_names": [
+    "MSE",
+    "MAE",
+    "RMSE",
+    "MAPE",
+    "sMAPE",
+    "MASE",
+    "ND",
+    "CRPS"
+  ],
+  "prediction_scale_factor": 1.0,
+  "model": "omni-scient-0.3.2",
+  "base_model": "amazon/chronos-2",
+  "finetuning": "LoRA (r=8), 1000 steps, on a small proprietary corpus of anonymized business time series (weekly demand series and high-frequency industrial sensor telemetry); no TIME data, no TIME upstream sources, no public benchmark corpora",
+  "checkpoint": "omni-scient-v0.3.2-lora"
+}

results/OmniScient/Coastal_T_S/20T/long/metrics.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b202df466e8cd746fc040f8b115840b6fe52d627380e2c008c4d1af08be70edd
+size 2479