trafo: add DRAM data sections to datagen

target/snitch_cluster/sw/apps/dnn/datagen.py

@@ -73,8 +73,9 @@ def emit_transformer_layer(name='transformer', **kwargs):
 
     ifmap = kwargs['ifmap']
     weights_q = kwargs['weights_q']
-    weights_k = kwargs['weights_k'].T
+    weights_k = kwargs['weights_k']
     weights_v = kwargs['weights_v']
+    weights_o = kwargs['weights_o']
 
     query = kwargs['query']
     key = kwargs['key']
@@ -104,10 +105,16 @@ def emit_transformer_layer(name='transformer', **kwargs):
     layer_str += f'\t.dtype = FP{kwargs["prec"]},\n'
     layer_str += '};\n\n\n'
 
+    # Declare the DRAM arrays
+    layer_str += f'static {dtype} {name}_Q_lin_dram[{S}][{E}] __attribute__((section(".data")));\n\n'
+    layer_str += f'static {dtype} {name}_K_lin_dram[{S}][{E}] __attribute__((section(".data")));\n\n'
+    layer_str += f'static {dtype} {name}_V_lin_dram[{S}][{E}] __attribute__((section(".data")));\n\n'
+    layer_str += f'static {dtype} {name}_O_dram[{S}][{P}] __attribute__((section(".data")));\n\n'
     layer_str += f'static {dtype} {name}_ifmap_dram[{S}][{E}] = ' + array_to_cstr(ifmap) + ';\n\n'
     layer_str += f'static {dtype} {name}_weights_q_dram[{E}][{P}] = ' + array_to_cstr(weights_q) + ';\n\n'
     layer_str += f'static {dtype} {name}_weights_k_dram[{E}][{P}] = ' + array_to_cstr(weights_k) + ';\n\n'
     layer_str += f'static {dtype} {name}_weights_v_dram[{E}][{P}] = ' + array_to_cstr(weights_v) + ';\n\n'
+    layer_str += f'static {dtype} {name}_weights_o_dram[{P}][{E}] = ' + array_to_cstr(weights_o) + ';\n\n'
     layer_str += f'static {dtype} {name}_query_dram[{S}][{P}] = ' + array_to_cstr(query) + ';\n\n'
     layer_str += f'static {dtype} {name}_key_dram[{S}][{P}] = ' + array_to_cstr(key) + ';\n\n'
     layer_str += f'static {dtype} {name}_value_dram[{S}][{P}] = ' + array_to_cstr(value) + ';\n\n'
@@ -921,16 +928,24 @@ def main():
                                 requires_grad=False, dtype=dtype)
         weights_v = torch.randn(param['input_dim']['embeddings'], param['input_dim']['positional_embeddings'],
                                 requires_grad=False, dtype=dtype)
+        weights_o = torch.randn(param['input_dim']['positional_embeddings'], param['input_dim']['embeddings'],
+                                requires_grad=False, dtype=dtype)
 
         query = transformer(ifmap, weights_q.T, None, None, None, False)
         key = transformer(ifmap, weights_k.T, None, None, None, False)
         value = transformer(ifmap, weights_v.T, None, None, None, False)
 
+        sm = torch.nn.Softmax(dim=-1)
+        attention = sm(torch.matmul(query, key.T))
+        attention_out = torch.matmul(attention, value)
+        transformer_out = torch.matmul(attention_out, weights_o)
+
         kwargs = {
             'ifmap': ifmap,
             'weights_q': weights_q,
             'weights_k': weights_k,
             'weights_v': weights_v,
+            'weights_o': weights_o,
             'query': query,
             'key': key,
             'value': value,