Upload learned parameters for llama3 in bit 8
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- params/llama3/8/fixed/woq/init/lm_head/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/lm_head/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.mlp.down_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.mlp.down_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.mlp.gate_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.mlp.gate_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.mlp.up_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.mlp.up_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.self_attn.k_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.self_attn.k_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.self_attn.o_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.self_attn.o_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.self_attn.q_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.self_attn.q_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.self_attn.v_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.0.self_attn.v_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.mlp.down_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.mlp.down_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.mlp.gate_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.mlp.gate_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.mlp.up_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.mlp.up_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.self_attn.k_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.self_attn.k_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.self_attn.o_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.self_attn.o_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.self_attn.q_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.self_attn.q_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.self_attn.v_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.1.self_attn.v_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.mlp.down_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.mlp.down_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.mlp.gate_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.mlp.gate_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.mlp.up_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.mlp.up_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.self_attn.k_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.self_attn.k_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.self_attn.o_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.self_attn.o_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.self_attn.q_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.self_attn.q_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.self_attn.v_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.10.self_attn.v_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.11.mlp.down_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.11.mlp.down_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.11.mlp.gate_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.11.mlp.gate_proj/zp.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.11.mlp.up_proj/scale.pt +0 -0
- params/llama3/8/fixed/woq/init/model.layers.11.mlp.up_proj/zp.pt +0 -0
params/llama3/8/fixed/woq/init/lm_head/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/lm_head/scale.pt and b/params/llama3/8/fixed/woq/init/lm_head/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/lm_head/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/lm_head/zp.pt and b/params/llama3/8/fixed/woq/init/lm_head/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.mlp.down_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.mlp.down_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.mlp.down_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.mlp.down_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.mlp.down_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.mlp.down_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.mlp.gate_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.mlp.gate_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.mlp.gate_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.mlp.gate_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.mlp.gate_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.mlp.gate_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.mlp.up_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.mlp.up_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.mlp.up_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.mlp.up_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.mlp.up_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.mlp.up_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.self_attn.k_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.k_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.k_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.self_attn.k_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.k_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.k_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.self_attn.o_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.o_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.o_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.self_attn.o_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.o_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.o_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.self_attn.q_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.q_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.q_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.self_attn.q_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.q_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.q_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.self_attn.v_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.v_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.v_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.0.self_attn.v_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.v_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.0.self_attn.v_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.mlp.down_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.mlp.down_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.mlp.down_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.mlp.down_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.mlp.down_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.mlp.down_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.mlp.gate_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.mlp.gate_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.mlp.gate_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.mlp.gate_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.mlp.gate_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.mlp.gate_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.mlp.up_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.mlp.up_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.mlp.up_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.mlp.up_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.mlp.up_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.mlp.up_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.self_attn.k_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.k_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.k_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.self_attn.k_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.k_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.k_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.self_attn.o_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.o_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.o_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.self_attn.o_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.o_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.o_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.self_attn.q_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.q_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.q_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.self_attn.q_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.q_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.q_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.self_attn.v_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.v_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.v_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.1.self_attn.v_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.v_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.1.self_attn.v_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.mlp.down_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.mlp.down_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.mlp.down_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.mlp.down_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.mlp.down_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.mlp.down_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.mlp.gate_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.mlp.gate_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.mlp.gate_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.mlp.gate_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.mlp.gate_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.mlp.gate_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.mlp.up_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.mlp.up_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.mlp.up_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.mlp.up_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.mlp.up_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.mlp.up_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.self_attn.k_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.k_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.k_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.self_attn.k_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.k_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.k_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.self_attn.o_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.o_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.o_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.self_attn.o_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.o_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.o_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.self_attn.q_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.q_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.q_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.self_attn.q_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.q_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.q_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.self_attn.v_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.v_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.v_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.10.self_attn.v_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.v_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.10.self_attn.v_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.11.mlp.down_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.11.mlp.down_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.11.mlp.down_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.11.mlp.down_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.11.mlp.down_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.11.mlp.down_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.11.mlp.gate_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.11.mlp.gate_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.11.mlp.gate_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.11.mlp.gate_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.11.mlp.gate_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.11.mlp.gate_proj/zp.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.11.mlp.up_proj/scale.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.11.mlp.up_proj/scale.pt and b/params/llama3/8/fixed/woq/init/model.layers.11.mlp.up_proj/scale.pt differ
|
|
params/llama3/8/fixed/woq/init/model.layers.11.mlp.up_proj/zp.pt
CHANGED
Binary files a/params/llama3/8/fixed/woq/init/model.layers.11.mlp.up_proj/zp.pt and b/params/llama3/8/fixed/woq/init/model.layers.11.mlp.up_proj/zp.pt differ
|
|