Searched refs:quantization (Results 1 – 12 of 12) sorted by relevance
25 * during quantization of gains. *28 #define nb_qua_gain6b 64 /* Number of quantization level */29 #define nb_qua_gain7b 128 /* Number of quantization level */
35 __u16 quantization; member
208 __u32 quantization; member1237 __u8 quantization; member
1189 __u32 quantization; member
644 * The quantization scale.653 * The quantization zero point.
4272 * * Quantized with symmetric per channel quantization for the filter:4279 * * Quantized signed with filter symmetric per channel quantization4970 * A version of quantized LSTM, using 16 bit quantization for internal4979 * cell. Tensor is quantized with a fixed quantization range of5052 * quantization range of [-2^4, 2^4 * 32767/32768] (scale = 2^4 /5057 * quantization range of [-1, 127/128] (scale = 1/128, zeroPoint =5064 * the current time step. Tensor is quantized using a quantization5069 * Tensor is quantized with a fixed quantization range of [-1, 127/128]5698 * * Quantized with symmetric per channel quantization for the filter:5711 * * Quantized signed with filter symmetric per channel quantization[all …]
304 * * Quantized with symmetric per channel quantization for the filter:316 * * Quantized signed with filter symmetric per channel quantization471 * * Quantized with symmetric per channel quantization for the filter:483 * * Quantized signed with filter symmetric per channel quantization3355 * * Quantized with symmetric per channel quantization for the filter:3361 * * Quantized signed with filter symmetric per channel quantization3965 * A version of quantized LSTM, using 16 bit quantization for internal3974 * cell. Tensor is quantized with a fixed quantization range of4047 * quantization range of [-2^4, 2^4 * 32767/32768] (scale = 2^4 /4611 * * Quantized with symmetric per channel quantization for the filter:[all …]
353 * * Quantized with symmetric per channel quantization for the filter:505 * * Quantized with symmetric per channel quantization for the filter:3163 * * Quantized with symmetric per channel quantization for the filter:3738 * A version of quantized LSTM, using 16 bit quantization for internal3747 * cell. Tensor is quantized with a fixed quantization range of3820 * quantization range of [-2^4, 2^4 * 32767/32768] (scale = 2^4 /3825 * quantization range of [-1, 127/128] (scale = 1/128, zeroPoint =3832 * the current time step. Tensor is quantized using a quantization3837 * Tensor is quantized with a fixed quantization range of [-1, 127/128]4365 * * Quantized with symmetric per channel quantization for the filter:[all …]
1064 * quantization unit
11000 NEUQUANT Neural-Net quantization algorithm by Anthony Dekker, 1994. See11001 "Kohonen neural networks for optimal colour quantization" in "Network:
1dictionary=main:en_us,locale=en_US,description=English (US),date ...
1dictionary=main:en,locale=en,description=English,date=1414726273, ...