…hub issue BVLC#3531)

The example talks about LevelDB as the db backend but has lmdb as the param in the execution.

This can happen if, e.g., testing never occurs in the log

This would've saved me an overnight download (slow connection here)

I tested it, and it worked for me.

refactor duplicate code into separate update function for smoothed loss

fix naming convention

…ype: boost::shared_ptr<caffe::Blob<float> >

and ScaleLayer.  The behavior of ChannelwiseAffineLayer can be
reproduced by a ScaleLayer with `scale_param { bias_term: true }`.

BiasLayer and ScaleLayer each take 1 or 2 bottoms, with the output having
the same shape as the first.  The second input -- either another  bottom or a
learned parameter -- will have its axes (virtually) broadcast and tiled to have
the same shape as the first, after which elementwise addition (Bias) or
multiplication (Scale) is performed.

The old uses `abspath`, which I think is so harmful:
* If I `cp -a` the whole project, `build/tools/caffe` still refer to
    the old file, until `make clean`, making debugging very hard
* For `tar` and `scp`, the soft links can not work
    unless the target project folder has the same path

…ile extracting libflags from absolute library path with multiple (dots)

The tests could not compile with USE_LEVELDB=0 and LevelDB missing from the system

`peformance` → `performance`

…toencoder. Arguments to matrix multiplication function are conditioned on this parameter, no actual transposing takes place.

test ip gradient computation with transpose on

for linking of the caffe tools, tests, etc. the library install name
needs to include the @rpath set for the executables and interfaces

this was broken on OSX by BVLC#3311

…). Replacing it with accessing Blob.shape[0] - for Blobs with num_axes() != 4

- add subheadings and list steps
- edit text, add comments, and try to make the code more understandable
- add new section for summary and encouragement to try your own image

- add subheadings and list steps for structure
- edit text and comments for clarity
- switch paths and use chdir for idempotency of scripts [shelhamer]
- title accuracy plots, rename ip -> fc to fit common naming, and
  rename output layer -> score [shelhamer]
- add experimentation section [shelhamer]

- add headings and text detail
- make nets by net spec
- define solvers inline through python protobuf
- do two-stage fine-tuning (first last layer alone, then end-to-end)
- show sample results

- create solvers inline through python protobuf
- drop manually written solver prototxt
- remove ordering prefix, since there is no real sequencing constraint
  for this example

This is important for the include order. Without this patch, a
previously installed caffe.pb.h might be included instead of the one
that is generated during the build.

Create an input layer to replace oddball Net `input` fields.

Drop special cases for `input` fields, the `Net` input members,
and the `Net` interface for Forward with bottoms along with
Forward() / ForwardPrefilled() distinction.

Restore the list of net inputs for compatibility with the
pycaffe and matcaffe interfaces and downstream C++.

Relax removal of `Forward()` variations by deprecating instead.

These can be used as direct replacements for the Caffe executable.

nvidia-docker requires long args with equal sign as of docker 1.10:
see BVLC#3518 (comment)

fixing lint errors

Use '//' instead of '/' for entire division.

…ication.

Fix some typos. Correct imports. Refactor data layers. Apply PEP8 formatting.

output info, warnings, and errors for fuller description of the upgrade

check all conditions all the time; V0 -> V1 and V1 -> V2 do not suffice.

convert inputs in legacy definitions (prototxt), but simply strip inputs
from legacy weights (caffemodel).

fix BVLC#3750

die loudly if a net definition (prototxt) mixes proto formats by
defining both `layer` and `layers` fields instead of complaining but
discarding and continuing.

fix BVLC#3381

- cosmetic change to mkl related doc

This is temporary measure to avoid an apparent upstream issue with
protobuf 3.0.0b2.post1.

This provides a framework for automatically aligning different layers of
a net despite up/downsampling, padding, and output size rounding.

- document by docstring and comment
- pep8
- add latest layers and alphabetize
- respect default crop params
- handle graphs with compositions of crops by walking only the
  first, cropped bottom of Crop layers
- make python3 happy by replacing arg tuple unpacking

- crop -> offset
- adjust crop axis by 1

- test known mappings: conv-pool-deconv stack, ReLU and 1x1 conv
- test effects of padding
- test rectangular/anisotropic coordinate mapping, test N-D
- catch error cases: negative crop, scale mismatch, tops that are not
  spatially connected

configure offset(s) through proto definition.

Changes are:
  reduce test blob dims for speed
  use standard Gaussian filler,
  polish formatting and rename tests,
  test HW crop and 5D crop,
  standard gradient checks

there seems to be a caching issue at the moment;
this is a temporary fix for BVLC#3786

MKLROOT variable is set by MKL scripts, so it also should be used in Makefile.

…ime on repeated calls.

…spec interface.

The deconvolution layer uses the convolution_param subgroup for its parameters, so hardcode that connection into param_name_dict().