Michel Pavlukhin

``` output_graph = "frozen_graph.pb"` output_graph_def = tf.graph_util.convert_variables_to_constants( sess, tf.get_default_graph().as_graph_def(), ["sample_sequence_2/while/Exit_3"] ) with tf.gfile.GFile(output_graph, "wb") as f: f.write(output_graph_def.SerializeToString()) ``` ```--------------------------------------------------------------------------- ValueError Traceback (most recent call last) in 8 # serialize and...

In paper authors did it, but I can't understand how they extracted latent codes from real images.

@TomeHirata @okhat I have a question about best design practices for dspy. Currently, dspy uses litellm to interact with models. If I want to integrate a compression model, it would...

I've made an implementation of TOON adapter for dspy Right now I'm running benchmarks. Maybe in future it could be merged into DSPy. https://github.com/Archelunch/dspy-toon