Decoder-Free Fully Transformer-based (DFFT) method is an efficient object detector based entirely on Decoder-Free Transformers. The Transformer backbone is focused on object detection. It extracts them at four scales and sends them to the next single-level density prediction module, dedicated to the encoder only. The prediction module first aggregates the multiscale object into a single object map using the Scale-Aggregated Encoder.
A lot of work has been done to implement the Decoder-Free Fully Transformer-based (DFFT) method by means of MQL5.
Testing of the work done:
Training and testing of the new model is carried out on the historical data of the symbol. Parameters of all indicators are used by default.To train the model, a huge number of random trajectories were collected on a time interval.The model turned out to be quite "light" in terms of consumption of computing resources both during training and in operation mode during testing.The training process was quite stable with smooth decrease of both Actor and Critic's error. The training process yielded a model capable of generating significant returns on both training and test data.
In conclusion, I would like to write that CyberNetic EA uses the latest trading algorithm, which gives us excellent results, both in live trading and in the tester startegies. Despite the fact that CyberNetic EA uses complex trading algorithms, it is very easy and simple to use, even for beginners.
