Time Series Forecasting
Transformers
Safetensors
Timer-S1
text-generation
time series
time-series
forecasting
foundation models
pretrained models
time series foundation models
quantized
4-bit precision
bitsandbytes
unofficial
custom_code
Instructions to use geetu040/Timer-S1-quantized-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use geetu040/Timer-S1-quantized-4bit with Transformers:
# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("geetu040/Timer-S1-quantized-4bit", trust_remote_code=True, dtype="auto") - Notebooks
- Google Colab
- Kaggle
Add files using upload-large-folder tool
Browse files- config.json +2 -2
config.json
CHANGED
|
@@ -37,7 +37,7 @@
|
|
| 37 |
"quantization_config": {
|
| 38 |
"_load_in_4bit": true,
|
| 39 |
"_load_in_8bit": false,
|
| 40 |
-
"bnb_4bit_compute_dtype": "
|
| 41 |
"bnb_4bit_quant_storage": "uint8",
|
| 42 |
"bnb_4bit_quant_type": "fp4",
|
| 43 |
"bnb_4bit_use_double_quant": false,
|
|
@@ -51,5 +51,5 @@
|
|
| 51 |
},
|
| 52 |
"rope_theta": 10000,
|
| 53 |
"transformers_version": "4.57.6",
|
| 54 |
-
"use_cache":
|
| 55 |
}
|
|
|
|
| 37 |
"quantization_config": {
|
| 38 |
"_load_in_4bit": true,
|
| 39 |
"_load_in_8bit": false,
|
| 40 |
+
"bnb_4bit_compute_dtype": "bfloat16",
|
| 41 |
"bnb_4bit_quant_storage": "uint8",
|
| 42 |
"bnb_4bit_quant_type": "fp4",
|
| 43 |
"bnb_4bit_use_double_quant": false,
|
|
|
|
| 51 |
},
|
| 52 |
"rope_theta": 10000,
|
| 53 |
"transformers_version": "4.57.6",
|
| 54 |
+
"use_cache": true
|
| 55 |
}
|