visdom not showing anything

[gari-marcos]

Hello, I configured the config.yaml to use visdom and although I start it in another terminal and I access the link in the browser I don't get any response. The train process gets to Setting up a new session... but the visdom window stays the same. Could you please help me?
Thank you so much!
This is the visdom log:

Checking for scripts.
It's Alive!
INFO:root:Application Started
You can navigate to http://localhost:8097
INFO:tornado.access:200 GET / (127.0.0.1) 19.52ms
INFO:tornado.access:200 GET /static/css/bootstrap.min.css?v=ec3bb52a00e176a7181d454dffaea219 (127.0.0.1) 4.04ms
INFO:tornado.access:200 GET /static/js/jquery.min.js?v=e071abda8fe61194711cfc2ab99fe104 (127.0.0.1) 2.87ms
INFO:tornado.access:200 GET /static/js/bootstrap.min.js?v=5869c96cc8f19086aee625d670d741f9 (127.0.0.1) 3.07ms
INFO:tornado.access:200 GET /static/css/react-resizable-styles.css?v=9f91a8dbf4d8f7ef1399e625660405f4 (127.0.0.1) 3.28ms
INFO:tornado.access:200 GET /static/css/react-grid-layout-styles.css?v=7dc8934d2f9ac5303b8f0bb1148152a0 (127.0.0.1) 3.37ms
INFO:tornado.access:200 GET /static/js/react-react.min.js?v=bca103da5b5404d93783ccf73e0e9d1e (127.0.0.1) 3.44ms
INFO:tornado.access:200 GET /static/js/react-dom.min.js?v=950495cc51ccb90612cf0fe0bb44f8f3 (127.0.0.1) 0.99ms
INFO:tornado.access:200 GET /static/js/layout_bin_packer.js?v=6c46683ed70fbb1443caf3531243836d (127.0.0.1) 1.82ms
INFO:tornado.access:200 GET /static/js/mathjax-MathJax.js?v=49565b9ce89c64da075a5a39969b366e (127.0.0.1) 2.20ms
INFO:tornado.access:200 GET /static/js/plotly-plotly.min.js?v=04dc3d6ed4e183246cd3ad70e259fcee (127.0.0.1) 12.04ms
INFO:tornado.access:200 GET /static/js/main.js?v=47fc3afa811b358b354df09184e392e7 (127.0.0.1) 17.40ms
INFO:tornado.access:200 GET /static/css/style.css?v=fb4729780f2360ef5d3dba13f8ff2873 (127.0.0.1) 17.76ms
INFO:tornado.access:200 GET /favicon.png (127.0.0.1) 9.99ms
INFO:tornado.access:200 GET /static/fonts/glyphicons-halflings-regular.woff2 (127.0.0.1) 0.68ms
INFO:tornado.access:101 GET /socket (127.0.0.1) 0.48ms
INFO:root:Opened new socket from ip: 127.0.0.1
INFO:tornado.access:200 POST /env/main (127.0.0.1) 0.90ms
INFO:tornado.access:200 POST /env/main (127.0.0.1) 0.80ms
INFO:tornado.access:200 GET /extensions/MathMenu.js?V=2.7.1 (127.0.0.1) 22.73ms
INFO:tornado.access:200 GET /extensions/MathZoom.js?V=2.7.1 (127.0.0.1) 36.97ms
INFO:tornado.access:200 POST /env/main (127.0.0.1) 0.99ms
INFO:tornado.access:101 GET /vis_socket (127.0.0.1) 0.94ms
INFO:root:Opened visdom socket from ip: 127.0.0.1
INFO:tornado.access:200 POST /env/main (127.0.0.1) 0.73ms

[MarcCote]

Hi @gari-marcos any success? I was able to get graphs displayed in visdom. For my test I ran
python train_action_prediction.py config.yaml where I changed some config settings (to speed up testing) as follows

diff --git a/config.yaml b/config.yaml
index de9054b..eb46dc3 100644
--- a/config.yaml
+++ b/config.yaml
@@ -1,9 +1,9 @@
 general:
-  task: "rl"  # cmd_gen, ap, sp, dgi, rl
+  task: "ap"  # cmd_gen, ap, sp, dgi, rl
   random_seed: 42
   use_this_many_data: -1
   use_cuda: True  # disable this when running on machine without cuda
-  visdom: False  # Plot training curves using Visdom
+  visdom: True  # Plot training curves using Visdom
 
   training:
     batch_size: 1  # for sp, change this to 64
@@ -18,13 +18,13 @@ general:
     patience: 3  # >=1 to enable
 
   evaluate:
-    run_eval: True
+    run_eval: False
     g_belief: False
     batch_size: 1  # for sp, change this to 64
     max_target_length: 100
 
   checkpoint:
-    report_frequency: 1000  # episode
+    report_frequency: 1 # episode
     save_frequency: 1000  # episode
     experiment_tag: 'exp_tag'
     load_pretrained: False  # during test, enable this so that the agent load your pretrained model
@@ -33,7 +33,7 @@ general:
     load_graph_update_model_from_tag: 'graph_update_model_tag'
 
   model:
-    use_pretrained_embedding: True
+    use_pretrained_embedding: False
     word_embedding_size: 300
     word_embedding_trainable: False
     node_embedding_size: 100

[MarcCote]

Also, make sure you select the main environment in visdom

[gari-marcos]

Thank you, it works now! I think I have some probelms with the config.yaml file. I think the error was at the use_pretrained_embedding parameter, as I had it setted to True.
Thank you again!

[MarcCote]

@gari-marcos
Note that we do use use_pretrained_embedding: True in our experiments but I disable it to speed up initialization (i.e. prevent loading crawl-300d-2M.vec.h5).
EDIT: It should work for both cases.