address an issue with triton auto-tuner and in-place calls. make the …

…assumption that after the first optimizer.step call, things are properly cached

lion_pytorch/lion_pytorch.py

@@ -47,6 +47,8 @@ def __init__(
         super().__init__(params, defaults)
 
         self.update_fn = update_fn
+        self.use_triton = use_triton
+        self.took_first_step = False
 
         if use_triton:
             from lion_pytorch.triton import update_fn as triton_update_fn
@@ -63,6 +65,13 @@ def step(
             with torch.enable_grad():
                 loss = closure()
 
+        # address an issue with autotune and in-place updates with triton
+        # on the first .step call, simply do not update parameters in-place, if using triton
+
+        update_kwargs = dict(inplace = False) if self.use_triton and not self.took_first_step else dict()
+
+        # update all parameters
+
         for group in self.param_groups:
             for p in filter(lambda p: exists(p.grad), group['params']):
 
@@ -82,7 +91,11 @@ def step(
                     lr,
                     wd,
                     beta1,
-                    beta2
+                    beta2,
+                    **update_kwargs
                 )
 
+        if not self.took_first_step:
+            self.took_first_step = True
+
         return loss

lion_pytorch/triton.py

@@ -1,4 +1,5 @@
 import torch
+from torch import Tensor
 
 try:
     import triton
@@ -7,6 +8,7 @@
     print('triton is not installed, please install by running `pip install triton -U --pre`')
     exit()
 
+# triton cuda kernel
 
 @triton.autotune(configs = [
     triton.Config({'BLOCK_SIZE': 128}, num_warps = 4),
@@ -72,19 +74,31 @@ def update_fn_kernel(
     tl.store(offset_exp_avg_ptr, exp_avg, mask = mask)
 
 def update_fn(
-    p: torch.Tensor,
-    grad: torch.Tensor,
-    exp_avg: torch.Tensor,
+    p: Tensor,
+    grad: Tensor,
+    exp_avg: Tensor,
     lr: float,
     wd: float,
     beta1: float,
-    beta2: float
+    beta2: float,
+    inplace: bool = True
 ):
     assert all([t.is_cuda for t in (p, grad, exp_avg)])
     n_elements = p.numel()
 
     grid = lambda meta: (triton.cdiv(n_elements, meta['BLOCK_SIZE']),)    
 
+    # address autotune and in-place update issue
+
+    if not inplace:
+        orig_p = p
+        orig_exp_avg = exp_avg
+
+        p = p.clone()
+        exp_avg = exp_avg.clone()
+
+    # call triton cuda kernel
+
     update_fn_kernel[grid](
         p,
         grad,
@@ -95,3 +109,9 @@ def update_fn(
         beta2,
         n_elements
     )
+
+    # update if not in-place call
+
+    if not inplace:
+        orig_p.copy_(p)
+        orig_exp_avg.copy_(exp_avg)

setup.py

@@ -3,7 +3,7 @@
 setup(
   name = 'lion-pytorch',
   packages = find_packages(exclude=[]),
-  version = '0.0.7',
+  version = '0.0.8',
   license='MIT',
   description = 'Lion Optimizer - Pytorch',
   author = 'Phil Wang',