Fix (scaling/standalone): better switch from runtime stats to param

src/brevitas/core/scaling/standalone.py

@@ -375,6 +375,12 @@ def __init__(
         self.restrict_scaling_pre = restrict_scaling_impl.restrict_init_module()
         self.restrict_threshold_pre = restrict_threshold_impl.restrict_init_module()
 
+    def init_scale(self):
+        if self.counter <= self.collect_stats_steps:
+            self.restrict_inplace_preprocess(self.buffer)
+            inplace_tensor_mul(self.value.detach(), self.buffer)
+            self.counter = self.collect_stats_steps + 1
+
     @brevitas.jit.script_method
     def training_forward(self, stats_input: Tensor, threshold: Tensor) -> Tensor:
         if self.counter < self.collect_stats_steps:
@@ -396,12 +402,10 @@ def training_forward(self, stats_input: Tensor, threshold: Tensor) -> Tensor:
             self.counter = new_counter
             return abs_binary_sign_grad(clamped_stats / threshold)
         elif self.counter == self.collect_stats_steps:
-            self.restrict_inplace_preprocess(self.buffer)
-            inplace_tensor_mul(self.value.detach(), self.buffer)
-            threshold = self.restrict_threshold(self.restrict_threshold_pre(threshold))
+            self.init_scale()
             value = self.clamp_scaling(self.restrict_scaling(self.value))
+            threshold = self.restrict_threshold(self.restrict_threshold_pre(threshold))
             value = value / threshold
-            self.counter = self.counter + 1
             return abs_binary_sign_grad(value)
         else:
             threshold = self.restrict_threshold(self.restrict_threshold_pre(threshold))

src/brevitas/graph/calibrate.py

@@ -65,7 +65,7 @@ def restore_return_quant_tensor(model, previous_state):
 def extend_collect_stats_steps(module):
     if hasattr(module, 'collect_stats_steps'):
         # We extend the collect steps in PTQ to match potentially long calibrations
-        module.collect_stats_steps = sys.maxsize
+        module.collect_stats_steps = sys.maxsize - 1
 
 
 def set_collect_stats_to_average(module):
@@ -80,6 +80,8 @@ def finalize_collect_stats(module):
         # otherwise the restrict_preprocess might be applied twice: during calibration
         # (that happens in training mode) and then when the model is evaluated
         module.counter = max(module.collect_stats_steps, module.counter)
+    if hasattr(module, 'init_scale'):
+        module.init_scale()
 
 
 class calibration_mode:

tests/brevitas/export/test_torch_qcdq.py

@@ -13,6 +13,7 @@
 
 @requires_pt_ge('1.9.1')
 @jit_disabled_for_export()
+@torch.no_grad()
 def test_torch_qcdq_wbiol_export(
         quant_module,
         quant_module_impl,
@@ -57,6 +58,7 @@ def test_torch_qcdq_wbiol_export(
 @requires_pt_ge('1.9.1')
 @jit_disabled_for_export()
 @parametrize('input_signed', [True, False])
+@torch.no_grad()
 def test_torch_qcdq_avgpool_export(input_signed, output_bit_width):
     in_size = (1, IN_CH, FEATURES, FEATURES)
     inp = torch.randn(in_size)