diff options
| -rw-r--r-- | nerv/examples/lmptb/tnn/layer_dag_t.lua | 43 | ||||
| -rw-r--r-- | nerv/examples/lmptb/unfold_ptb_main.lua | 9 |
2 files changed, 29 insertions, 23 deletions
diff --git a/nerv/examples/lmptb/tnn/layer_dag_t.lua b/nerv/examples/lmptb/tnn/layer_dag_t.lua index 1a89816..cd5fba7 100644 --- a/nerv/examples/lmptb/tnn/layer_dag_t.lua +++ b/nerv/examples/lmptb/tnn/layer_dag_t.lua @@ -22,6 +22,7 @@ local function discover(id, layers, layer_repo) local layer = layer_repo:get_layer(id) local dim_in, dim_out = layer:get_dim() ref = { + id = layer.id, layer = layer, inputs = {}, outputs = {}, @@ -51,26 +52,33 @@ function DAGLayerT:__init(id, global_conf, layer_conf) local ref_from = discover(id_from, layers, layer_conf.sub_layers) local ref_to = discover(id_to, layers, layer_conf.sub_layers) local input_dim, output_dim, _ - if ref_from and ref_from.outputs[port_from] ~= nil then - nerv.error("%s has already been attached", from) - end - if ref_to and ref_to.inputs[port_to] ~= nil then - nerv.error("%s has already been attached", to) - end if id_from == "<input>" then input_dim, _ = ref_to.layer:get_dim() if dim_in[port_from] ~= input_dim[port_to] then nerv.error("mismatching data dimension between %s and %s", from, to) end inputs[port_from] = {ref_to, port_to} - ref_to.inputs[port_to] = inputs -- just a place holder + if ref_to.inputs[1] == nil then + ref_to.inputs[1] = {} + end + if ref_to.inputs[1][port_to] ~= nil then + nerv.error("port(%d) for layer(%s) already attached", port_to, to) + end + ref_to.inputs[1][port_to] = inputs -- just a place holder elseif id_to == "<output>" then _, output_dim = ref_from.layer:get_dim() if output_dim[port_from] ~= dim_out[port_to] then nerv.error("mismatching data dimension between %s and %s", from, to) end outputs[port_to] = {ref_from, port_from} - ref_from.outputs[port_from] = outputs -- just a place holder + if ref_from.outputs[1] == nil then + ref_from.outputs[1] = {} + end + if ref_from.outputs[1][port_from] ~= nil then + nerv.error("port(%d) for layer(%s) already attached", port_from, from) + end + ref_from.outputs[1] = {} + ref_from.outputs[1][port_from] = outputs -- just a place holder else _, output_dim = ref_from.layer:get_dim() input_dim, _ = ref_to.layer:get_dim() @@ -134,10 +142,10 @@ function DAGLayerT:__init(id, global_conf, layer_conf) end function DAGLayerT:init(batch_size, chunk_size) - nerv.printf("initing DAGLayerT\n") + nerv.info("initing DAGLayerT...\n") if chunk_size == nil then chunk_size = 1 - nerv.printf("(Initing DAGLayerT) chunk_size is nil, setting it to default 1\n") + nerv.info("(Initing DAGLayerT) chunk_size is nil, setting it to default 1\n") end self.chunk_size = chunk_size @@ -179,12 +187,12 @@ function DAGLayerT:init(batch_size, chunk_size) end for id, ref in pairs(self.layers) do for i = 1, ref.input_len do - if ref.inputs[i] == nil then + if ref.inputs[1][i] == nil then --peek at time 1 nerv.error("dangling input port %d of layer %s", i, id) end end for i = 1, ref.output_len do - if ref.outputs[i] == nil then + if ref.outputs[1][i] == nil then --peek at time 1 nerv.error("dangling output port %d of layer %s", i, id) end end @@ -258,6 +266,7 @@ function DAGLayerT:set_inputs(input, t) end local layer = self.inputs[i][1] local port = self.inputs[i][2] + layer.inputs[t][port] = input[i] end end @@ -296,10 +305,8 @@ function DAGLayerT:update(bp_err, input, output, t) self:set_err_inputs(bp_err, t) self:set_inputs(input, t) self:set_outputs(output, t) - -- print("update") for id, ref in pairs(self.queue) do - -- print(ref.layer.id) - ref.layer:update(ref.err_inputs, ref.inputs, ref.outputs, t) + ref.layer:update(ref.err_inputs[t], ref.inputs[t], ref.outputs[t], t) end end @@ -312,8 +319,7 @@ function DAGLayerT:propagate(input, output) local ret = false for i = 1, #self.queue do local ref = self.queue[i] - -- print(ref.layer.id) - ret = ref.layer:propagate(ref.inputs, ref.outputs, t) + ret = ref.layer:propagate(ref.inputs[t], ref.outputs[t], t) end return ret end @@ -328,8 +334,7 @@ function DAGLayerT:back_propagate(bp_err, next_bp_err, input, output, t) self:set_outputs(output, t) for i = #self.queue, 1, -1 do local ref = self.queue[i] - -- print(ref.layer.id) - ref.layer:back_propagate(ref.err_inputs, ref.err_outputs, ref.inputs, ref.outputs, t) + ref.layer:back_propagate(ref.err_inputs[t], ref.err_outputs[t], ref.inputs[t], ref.outputs[t], t) end end diff --git a/nerv/examples/lmptb/unfold_ptb_main.lua b/nerv/examples/lmptb/unfold_ptb_main.lua index 6c4ead3..5affadf 100644 --- a/nerv/examples/lmptb/unfold_ptb_main.lua +++ b/nerv/examples/lmptb/unfold_ptb_main.lua @@ -8,6 +8,7 @@ require 'lmptb.lmvocab' require 'lmptb.lmfeeder' require 'lmptb.lmutil' +require 'tnn.init' nerv.include('lmptb/layer/init.lua') --[[global function rename]]-- @@ -70,7 +71,7 @@ function prepare_layers(global_conf, paramRepo) }, ["nerv.SelectLinearLayer"] = { - ["selectL1"] = {{["ltp"] = "ltp_ih"}, {["dim_in"] = {1}, ["dim_out"] = {global_conf.hidden_size}}}, + ["selectL1"] = {{["ltp"] = "ltp_ih"}, {["dim_in"] = {1}, ["dim_out"] = {global_conf.hidden_size}, ["vocab"] = global_conf.vocab}}, }, ["nerv.SigmoidLayer"] = { @@ -90,7 +91,7 @@ function prepare_layers(global_conf, paramRepo) for i = 1, global_conf.bptt do layers["nerv.IndRecurrentLayer"]["recurrentL" .. (i + 1)] = recurrentLconfig layers["nerv.SigmoidLayer"]["sigmoidL" .. (i + 1)] = {{}, {["dim_in"] = {global_conf.hidden_size}, ["dim_out"] = {global_conf.hidden_size}}} - layers["nerv.SelectLinearLayer"]["selectL" .. (i + 1)] = {{["ltp"] = "ltp_ih"}, {["dim_in"] = {1}, ["dim_out"] = {global_conf.hidden_size}}} + layers["nerv.SelectLinearLayer"]["selectL" .. (i + 1)] = {{["ltp"] = "ltp_ih"}, {["dim_in"] = {1}, ["dim_out"] = {global_conf.hidden_size}, ["vocab"] = global_conf.vocab}} end local layerRepo = nerv.LayerRepo(layers, paramRepo, global_conf) printf("%s preparing layers end.\n", global_conf.sche_log_pre) @@ -138,7 +139,7 @@ function prepare_dagLayer(global_conf, layerRepo) printf("\t%s->%s\n", key, value) end - local dagL = nerv.DAGLayer("dagL", global_conf, {["dim_in"] = dim_in_t, ["dim_out"] = {1}, ["sub_layers"] = layerRepo, + local dagL = nerv.DAGLayerT("dagL", global_conf, {["dim_in"] = dim_in_t, ["dim_out"] = {1}, ["sub_layers"] = layerRepo, ["connections"] = connections_t, }) dagL:init(global_conf.batch_size) @@ -277,7 +278,7 @@ if (set == "ptb") then global_conf = { lrate = 1, wcost = 1e-6, momentum = 0, cumat_type = nerv.CuMatrixFloat, - mmat_type = nerv.CuMatrixFloat, + mmat_type = nerv.MMatrixFloat, hidden_size = 200, batch_size = 10, |