Spaces:
Running
Running
Commit
·
a13918e
1
Parent(s):
86a28a0
Remove \cross for an \ast
Browse files- config/constants.py +1 -1
- data_processing.py +2 -2
- handlers/leaderboard_handlers.py +1 -1
- static/html_content.py +1 -1
config/constants.py
CHANGED
|
@@ -18,7 +18,7 @@ DISCARDED_MODELS = {
|
|
| 18 |
|
| 19 |
}
|
| 20 |
|
| 21 |
-
TASKS = ["Spec-to-RTL", "Code Completion", "Line Completion
|
| 22 |
S2R_BENCHMARKS = ["VerilogEval S2R", "RTLLM"]
|
| 23 |
CC_BENCHMARKS = ["VerilogEval MC", "VeriGen"]
|
| 24 |
LC_BENCHMARKS = ["RTL-Repo"]
|
|
|
|
| 18 |
|
| 19 |
}
|
| 20 |
|
| 21 |
+
TASKS = ["Spec-to-RTL", "Code Completion", "Line Completion *", "Module Completion"]
|
| 22 |
S2R_BENCHMARKS = ["VerilogEval S2R", "RTLLM"]
|
| 23 |
CC_BENCHMARKS = ["VerilogEval MC", "VeriGen"]
|
| 24 |
LC_BENCHMARKS = ["RTL-Repo"]
|
data_processing.py
CHANGED
|
@@ -55,7 +55,7 @@ def filter_leaderboard(task, benchmark, model_type, search_query, max_params, st
|
|
| 55 |
valid_benchmarks = CC_BENCHMARKS
|
| 56 |
if benchmark == "All":
|
| 57 |
subset = subset[subset["Benchmark"].isin(valid_benchmarks)]
|
| 58 |
-
elif task == "Line Completion
|
| 59 |
valid_benchmarks = LC_BENCHMARKS
|
| 60 |
if benchmark == "All":
|
| 61 |
subset = subset[subset["Benchmark"].isin(valid_benchmarks)]
|
|
@@ -89,7 +89,7 @@ def filter_leaderboard(task, benchmark, model_type, search_query, max_params, st
|
|
| 89 |
return filter_bench_all(subset, state.get_current_agg(), agg_column="Agg S2R", name=name)
|
| 90 |
elif task == "Code Completion":
|
| 91 |
return filter_bench_all(subset, state.get_current_agg(), agg_column="Agg MC", name=name)
|
| 92 |
-
elif task == "Line Completion
|
| 93 |
return filter_RTLRepo(subset, name=name)
|
| 94 |
elif benchmark == "RTL-Repo":
|
| 95 |
return filter_RTLRepo(subset, name=name)
|
|
|
|
| 55 |
valid_benchmarks = CC_BENCHMARKS
|
| 56 |
if benchmark == "All":
|
| 57 |
subset = subset[subset["Benchmark"].isin(valid_benchmarks)]
|
| 58 |
+
elif task == "Line Completion *":
|
| 59 |
valid_benchmarks = LC_BENCHMARKS
|
| 60 |
if benchmark == "All":
|
| 61 |
subset = subset[subset["Benchmark"].isin(valid_benchmarks)]
|
|
|
|
| 89 |
return filter_bench_all(subset, state.get_current_agg(), agg_column="Agg S2R", name=name)
|
| 90 |
elif task == "Code Completion":
|
| 91 |
return filter_bench_all(subset, state.get_current_agg(), agg_column="Agg MC", name=name)
|
| 92 |
+
elif task == "Line Completion *":
|
| 93 |
return filter_RTLRepo(subset, name=name)
|
| 94 |
elif benchmark == "RTL-Repo":
|
| 95 |
return filter_RTLRepo(subset, name=name)
|
handlers/leaderboard_handlers.py
CHANGED
|
@@ -38,7 +38,7 @@ def create_leaderboard_handlers(
|
|
| 38 |
elif task == "Code Completion":
|
| 39 |
new_benchmarks = ["All"] + CC_BENCHMARKS
|
| 40 |
new_simulators = SIMULATORS
|
| 41 |
-
elif task == "Line Completion
|
| 42 |
new_benchmarks = LC_BENCHMARKS
|
| 43 |
new_simulators = SIMULATORS
|
| 44 |
elif task == "Module Completion":
|
|
|
|
| 38 |
elif task == "Code Completion":
|
| 39 |
new_benchmarks = ["All"] + CC_BENCHMARKS
|
| 40 |
new_simulators = SIMULATORS
|
| 41 |
+
elif task == "Line Completion *":
|
| 42 |
new_benchmarks = LC_BENCHMARKS
|
| 43 |
new_simulators = SIMULATORS
|
| 44 |
elif task == "Module Completion":
|
static/html_content.py
CHANGED
|
@@ -81,7 +81,7 @@ INTRO_HTML = """
|
|
| 81 |
|
| 82 |
LC_FOOTNOTE_HTML = """
|
| 83 |
<div id="lc-footnote" style="font-size: 13px; opacity: 0.6; margin-top: -5px; z-index:999; text-align: left;">
|
| 84 |
-
<span style="font-weight: 600; opacity: 1;"
|
| 85 |
<em>Line Completion</em> excludes "reasoning" models since this task targets quick auto-completion<br/>
|
| 86 |
Additionally, for <em>Line Completion</em> and <em>Code Completion</em> benchmarks we use <b>Base</b> model variant (if available), and for <em>Spec-to-RTL</em> we use <b>Instruct</b> model variant
|
| 87 |
</div>
|
|
|
|
| 81 |
|
| 82 |
LC_FOOTNOTE_HTML = """
|
| 83 |
<div id="lc-footnote" style="font-size: 13px; opacity: 0.6; margin-top: -5px; z-index:999; text-align: left;">
|
| 84 |
+
<span style="font-weight: 600; opacity: 1;">*</span>
|
| 85 |
<em>Line Completion</em> excludes "reasoning" models since this task targets quick auto-completion<br/>
|
| 86 |
Additionally, for <em>Line Completion</em> and <em>Code Completion</em> benchmarks we use <b>Base</b> model variant (if available), and for <em>Spec-to-RTL</em> we use <b>Instruct</b> model variant
|
| 87 |
</div>
|